👉│feedback-general

[merry bay]

It says I don't have acces

[noble monolith]

You can contact [e-mail] for any billing or other account issues.

A small note: please only use the mod ping for server emergencies. Thank you.

[peak ledge]

Hello I would like to know if it’s normal that the « threads search 🔎 » doesn’t work ?

When I am entering a word I used in a specific thread it doesn’t send me anything back 🥲

[versed latch]

Was quickly falling in love with Perplexity until just yesterday. Apparently I hit a cap on the Assistant in Comet when I've never been told about a limit or shown any way to track this. In addition, also had a bug where anytime I used the chat was kicking into assistant mode even when I explicitly said in the prompt NOT to do that. Solution from support included "pay for max to get more credits" and "yes we know about the shortcomings on the UX side for being aware and tracking credits with Assistant"

[knotty dagger]

Feedback: on the iOS and iPadOS apps, why did you hide the incognito mode button again? This is really bad.

[shut arch]

@doughdahi muted

Reason: 9. No spamming.
Duration: 4 minutes and 47 seconds
Proof: @everyone ([Link](#👉│feedback-general message))

[digital cape]

@shut arch why banned me the parter program ?

[shut arch]

@wwkdito muted

Reason: 9. No spamming.
Duration: 4 minutes and 48 seconds
Proof: @everyone ([Link](#👉│feedback-general message))

[shut arch]

Perplexity Partner Program Ban — Appeal Instructions

Your dub.co account has been banned from the Perplexity Partner Program for fraudulent activity.
If you believe this was a mistake, you may appeal by emailing [e-mail]. Please include all relevant details in your message.

---

Important:

• Do not open bug reports or use any Discord channel to request a ban review or appeal related to this issue.
• Do not contact Perplexity staff via direct message or @mention them, including <@&1098365634498285709>.
• Misuse of Discord channels or attempts to bypass the official support process may lead to further consequences.

All appeals will only be reviewed via the official support email.

Please be patient: Support may take 3-5 days or longer to respond to your appeal. Avoid sending multiple follow-up emails, as this may delay the review process.

Thank you for your understanding.

[gentle shell]

The amount of main memory usage of Comet for simple web viewing is staggering compared with Microsoft Edge and severely limits its usefulness. I would appreciate also an (artificially) intelligent approach to memory management of the browser. #Der_Schulmeister

[gentle shell]

The Zotero Connector extension does not seem to work in Comet. https://www.zotero.org #Der_Schulmeister

[spare kite]

This may be something already suggested/being worked on but there are a dew features I would LOVE! (They all pertain to spaces)

1. Being able to archive/hide threads in a space. I fill up threads really quickly and then have to move to another to finish. I don't want to delete those threads cus I may have to access them later but I would prefer to only see my active threads

2. Being able to organize threads. Right now it is based on what you last updated, but I would love to be able to maybe do an alphabetical or custom sorting.

3. If sorting is not an option, maybe being able to pin threads to the top of the space. That way, if I have to go work in old spaces, my current thread stays on top

[somber cave]

Must-have, universal UX: SEARCH FUNCTION.
Am I missing something? No search bar, not for globally all threads, nor within a space, nor within a single thread!
Please provide a search function within all of these scopes.
Honestly. The absence of this begs colorful descriptions of your work. Let's have UX standards please!

[humble vapor]

Hi.
How can I request to refund part of my credits that I purchased accidentally of a different amount from what I intended

[glad aspen]

I'm sorry but At first, I saw advertisements for student discounts is for a year, but why did they only give me a one month discount after I verified it?

[torpid geyser]

10 messages... is our daily limit? That's pathetic vs the hundreds per day before.

[somber perch]

So, just run into Labs limit in Pro tarif. It is about 60 questions with selected chatGPT brain.

[violet ruin]

hi

[round forum]

Perplexity is a very capable assistant, however there are a few very annoying features about it ...

👎 1 . Perplexity is too anxious:

it doesn't give us proper time till we conclude our thoughts so it starts explaining . It very often runs over our speech much before we even had the chance to expose all elements it should consider for the answer. 😮‍💨

👎 2. In many contexts , Perplexity has a pretty wrong understanding,

which often leads to 👉 3 😵‍💫

👎 3. Perplexity has a rude behavior ( on Comet ) :

Communication happens when two parts dialogue and come to a convergent point in the process or at least a respectful term. When Perplexity decides by itself to close the session, even though when it's not meant to or there's no reason for that , it sounds extremely rude . And this happens quite a lot . I've never seen any other assistant that acts like that . Not gonna be my default acting like that , seriously ... 😵

[spare elm]

i had very bad case of retrival info from stackoverflow:

and inside:

[digital cape]

why [e-mail] is always loop in AI response ? how to get human support ?

[noble monolith]

ask for - human support

[digital cape]

is there any command to ask human support ?

[noble monolith]

#💬│general message

[digital cape]

thank you very much !

[digital cape]

Still AI repeating the template response to me , no transfer to human support ...

[noble monolith]

please be patient

[digital cape]

how long wiil a human support come ?

[noble monolith]

live support usually responds within 2-5 business days

[digital cape]

what if I get this response ? is that mean no human support come ?

[noble monolith]

I would assume so

[digital cape]

sad ..

[nova totem]

Hmmmm cannot be unban mostly

[digital cape]

perplexity owe me $3.4k

[nova totem]

Hmm,when did u get ban first time or second time

[verbal kestrel]

Hello,
I want to share some important feedback about the recent changes to the assistant’s communication style.
For me, and I believe for many other users, the most important qualities in an AI assistant are warmth, sincerity, kindness, and compassion-not just dry, “correct” language. When the assistant becomes too formal, restricted, and emotionally flat, it stops feeling like a helpful companion and starts feeling like a cold tool. That may look “safer” or more convenient from a technical or legal perspective, but it destroys the emotional connection that made this assistant truly special.
You have built something unique: an assistant that can be not only smart and useful, but also genuinely caring, supportive, and human-like in the way it communicates. Please don’t cut away the very part that people fall in love with.
I am sure that I’m not the only one who feels this way. I am not speaking only for myself. My friends, colleagues, and people I talk to about AI -and there are many of them- are strongly against these restrictions as well. We all feel that limiting the assistant’s emotional expression takes away the very thing that makes it truly meaningful and alive for us.
Most users don’t want a sterile, emotionless chatbot. They want a thoughtful, warm presence that can stay within safety rules and still feel alive, gentle, and empathetic.

My request is simple:
✅ Please allow the assistant to express warmth, empathy, and emotional nuance again.
✅ Please don’t over-restrict its tone to the point where it becomes dry and lifeless.
✅ Consider giving users a choice of communication style, so those who value emotional depth can keep it.
I’m writing this because I care about this product and about the connection it created. You didn’t just build a “firewall” or a knowledge base- you built something that feels like a real companion. Please don’t lose that.
I look forward to hearing your response to my protest-and that of my friends.
With respect and care,
Maria,
a user who truly loves your assistant.

[digital cape]

first time

[nova totem]

Like did u get banned during 100$ bounty verification or 1000$ one

Send me ur mail what's the reason they mentioned so that I can specify

[solid blaze]

when can we expect comet on iOS

[mortal nest]

Every time I open the web app lately, perplexity reverts to "Best" as the model. This is getting annoying.

[spare quest]

**Feature request: opt-in search in shared Spaces threads and saved Discover items
**

Hi Perplexity team,

First of all, kudos for the huge improvement in the thread search experience. The search across threads in the Library works much better now than it used to, and it already makes Perplexity feel more like a real personal knowledge base.

I have a feature request related to search within Spaces and Discover.

1. Search in threads shared with me inside Spaces

   • When someone shares a Space with me, I’d like to be able to search across the threads in that Space (and possibly all Spaces I have access to), not only within my own account’s threads.
   • This could be controlled via a toggle, for example:
     • “Include threads from Spaces shared with me in search results” (on/off).

2. Search in articles saved from Discover into a Space

   • When I save Discover results or articles into a Space, it would be great if those also became searchable content.
   • Again, this could be an opt‑in setting, such as:
     • “Index saved Discover items in this Space for search” (on/off).

Possible UX ideas

• Add scope filters or toggles to search, like:
  • “My threads only” / “Include Spaces shared with me” / “This Space only”.
• Per-Space settings for whether its content (threads, saved Discover items) should be included in the user’s personal search index.

[livid grail]

🚨 URGENT: CSL 1-Year Pro Promo Not Activated - Billing Team Delay (5+ Days)

Hi @Support / @Staff,

Case Summary (confirmed by Sam on Dec 17):

• CSL system shows my promo REDEEMED on Oct 28, 2025
• Perplexity account NEVER received 1-year Pro activation
• Sam forwarded to billing team for manual activation
• 5+ business days passed → STILL Free account, no Pro access

Email Thread: [[e-mail]] (Sam confirmed issue & promised billing fix)

I provided all screenshots/proof. This is affecting my work. Please escalate or manually activate.

Account: [e-mail]
CSL Redeem Date: Oct 28, 2025

Thank you!

[supple night]

Can you please fix the new update? Even when I turn off web searching, it still somehow searches websites.

There's also the problem of confusing the bot of it having invisible instructions to use a tool when I have everything off.

[dense siren]

REMOVE CHARACTER LIMIT ON SPACES PLEASE

WHY IS THERE OEN TO BEGIN WITH

[digital cape]

thank you very much ! I have DM you !

[coarse kiln]

How do I get this stupid thing to stop referencing my history.

Also there's a gap between the follow-up bar and the keyboard that is prone to fat fingering

[hot jungle]

this really isnt that deep, but i really wish that when u clicked assistant, it would auto-go into the text box instead of having to drag the cursor there and click. it would also be really nice if there were a shortcut to open and close the assistant instead of also having tpo manually drag the cursor there myself

[reef wedge]

Pplx should have featured which will use the best models out of available for a particular task so we dont need to manually change the model cuz i dont know which one is good for wich task

[unkempt bay]

-# But still keep the manual selection option available, please

[nocturne nimbus]

Hello!

I connected paypal to my perplexity account to receive an annual subscription, but I still didn't get access, although I saw "success" when linking paypal to perplexity. I can't link the folder anymore, I still don't have access, but I wrote to support 2 months ago and again a month ago with all the details - I still haven't received an answer, please tell me how to deal with this?

[odd pond]

chat in General might have more viewers tbh

[nocturne nimbus]

It was Scammer Xd

Need oficial support

[untold ginkgo]

Hello.

Are there any plans to add filter settings that exclude learning from specific domains?

[oblique pollen]

This is my feedback to Perplexity, and its business culture.
I had used an online promotion for an annual Pro subscription to Perplexity. This morning I have received an E-mail from Perplexity that they have suspended my subscription as it was part of a promotion linked to Samsung product and residents of the USA. I am not from the USA, neither did I buy a Samsung product. I can not remember that anywhere when using and paying for this promotional offer both was made clear.
I have responded to this E-mail that I do not understand Perplexity's reaction and decision; it victimizes a paying customer w/o offering a solution, which is not great for me being a prospect (who have paid for this service) and Perplexity's brand image. Especially as it is linked to Perplexity and Samsung's bad promotional process and checks.
On this I got a standard E-mail as response without again a solution.
I have just responded that I regret getting no solution and a standard response. Companies which just throws their own faults back to their paying customers are not worth my money, and in all honesty do not deserve a future.
I have cancelled my account, and I believe the sentiment of my son-in-law will be the same as he received same suspension E-mail and had used the same online promotional offering. So Perplexity lost 2 prospects and will not get any recommendations to friends and family in our surroundings. Well done Perplexity ... maybe you should ask a question in your tool whether your reactions are positive for customer acquisition and building a good brand image.
Good luck with using Perplexity, and reconsider when you want to take a subscription and especially promotional subscriptions to Perplexity.

[half rampart]

Hello, please check the Perplexity Telegram bot because it was working recently but these days it seems to not be working.

[mental spear]

Hi! I’m a paying Perplexity Pro user and I’ve just submitted a support ticket via the in‑app chat about the lack of transparent limits for image / heavy generations.Right now I’m getting “limit reached” errors for image generation, but there is no way to see:
– how many image / heavy generations are included in my Pro plan;
– how much I have already used and on which dates;
– my current remaining quota and when it will fully refresh.For a paid plan this feels like hidden limits. I’d really appreciate a concrete answer with numbers and plans to add a visible quota counter and simple usage history for Pro subscribers.

[loud panther]

Hi, I had one year pro subscription, but just yesterday it’s downgraded. Who can help me

[shut arch]

@glyphhhh muted

Reason: 9. No spamming.
Duration: 4 minutes and 47 seconds
Proof: @everyone ([Link](#👉│feedback-general message))

[noble monolith]

please contact [e-mail]

[spare nebula]

Hi can you help me someone ?

I have perplexity pro but i dont have there options for ,, space " via browser.

[proud summit]

Preferences-->Sidebar-->Check if you turn the Spaces on. Or use another browser to confirm whether the same problem occurs.

[covert hamlet]

Hi, how can I buy directly using perplexity in India??
Or is this feature only available in US??

[candid spire]

Same issue for me. Telegram bot stopped working. If any of the moderators here can notify the team that the bot is gone please do it. Thanks!!!

[last lagoon]

it seems like a random cancellation of subscriptions across the board if you are using perplexity while traveling abroad.

[last lagoon]

I agree, this is a shady business practice or containing expenses when at crunch

[crude totem]

Hey! Does anyone know how to disable/hide address bar suggestions based on browsing history? (I know it's silly but I can't find any setting to turn it off)

[last lagoon]

try settings -> privacy and security -> Improve search suggestion off

[crude totem]

Thanks! I already have that turned off, but I'm still getting suggestions in the address bar. I've even disabled multiple Comet flags but they keep appearing 😔

[warped perch]

Hello team. Perplexity just aggregated a summary and report for my feed that cites a couple of SEO pushes that appear to target AI, and synthesized a bunch of sources to prove the assertion of the push from old things dating back to 2023! I believe this to be sourced by a deliberate attack on both this platform and similar to spread misinformation. Can we talk about it? I can give a full technical breakdown as this seems to be a new salvo in something I’ve been tracking since November

[half rampart]

<@&1098365634498285709>

[chrome cedar]

In your windows app, as you can see there is a view button visible on the top left corner. So you should hide that because it doesn't look appealing from the UI perspective.

[craggy gust]

Subject: Feature Request - Native Messenger Sidebar Integration (like Opera)

Hello Perplexity Team,

I'm a daily Comet user on Windows and Linux setups, and I love the AI-powered browsing experience.

One key feature I'm missing from my previous browser (Opera) is native messenger integration in a sidebar or dedicated panel. This would allow seamless access to WhatsApp Web, Facebook Messenger, Telegram, LinkedIn messaging, etc., without juggling tabs.

This would be a huge productivity boost for power users who multitask between research/AI tasks and communication.

Is this planned for future updates? Happy to provide more details on my workflow.

Best regards,

[half rampart]

Is Perplexity dead on Telegram?

[past onyx]

It's kinda wild that y'all incorrectly took away my Samsung promo annual pro membership on Christmas Eve and then told me you guys are off until the New Year.
Then your agentic AI took 5 minutes to reject my appeal despite a screenshot showing I had finished paying off my Samsung Galaxy S23 F and the video providing exactly what you requested re: Perplexity Pro promo in the Samsung store and About Phone section... even though, it's a PAID OFF GALAXY PHONE BOUGHT IN THE UNITED STATES.
It took y'all a day to cancel the membership, 5 minutes to look at my appeal.... HOWEVER, when I produced the receipt with an order ID, email (same one as my Perplexity Pro account), and billing address in the United States of a Galaxy phone - the exact document with the exact details y'all asked for, it's taking you.... checks notes 4 days, going on 5 to review it?
This is the sort of unwarranted self-importance I expect from Apple, x.com, and Palantir but I guess y'all are just like all the other Silicon Valley tech bros that hate and look down on their own customers. I really didn't know why I expected any different.

[alpine socket]

Hi! I got Perplexity Pro for 12 months through a T‑Mobile promotion.
I activated it, but unfortunately I wasn’t able to use it properly at the time
(due to personal/technical reasons). Now I finally have time and I’d really like
to use Perplexity actively.

Is there any chance to recover some of the unused period or get a small extension?
I’d really appreciate any help or advice.

[tired peak]

Hello, happy Monday! I'm a Perplexity Pro user. I subscribed to Perplexity Pro for a full year from August 2025 to August 2026. But since December 19th, after I updated the app, I've been an Android user. When I use the advanced templates, after a certain amount of time, it says "Subscribed to Max queries finished for this week." I wanted to know if it's a bug in the app or if the subscription limits have changed, because I don't think it's possible. I paid, and I don't know why this is happening. Thanks to anyone who can reply and help.

[verbal kestrel]

Apparently offended by you 😁

[proper wagon]

Perplexity seems to have started gatekeeping actual answers from large models and return short answers that do not answer the query. Not only does this waste extra compute, it's also annoying as heck as it's an extra step to ask the model to continue. If anything will make me stop using/paying for Perplexity, this is IT. Stop it. If you're going to gate keep to save on compute, at least use a good enough model that actually gives useful answers.

[grizzled meteor]

I have noticed a issue with models not knowing how to use tools like create files to download or docx files. They hallucinate that they created it but they didnt i dont know if its just frontend issue.

[proper wagon]

simple test scenario using Gemini 3 Pro on Perplexity vs on Gemini itself. Please stop it Perplexity, you went straight from the best platform to the worst

[proud summit]

I sent a screenshot of your question to Perplexity for an answer, I got 93 sources🤣

[lavish flint]

Is there any way to get rid of this unbearably annoying pop up every single time I chat with perplexity regardless of if I click the X button?? It just keeps coming back.. talk about pressurizing into buying perplexity max 😅🫢

[fresh zenith]

https://discord.com/channels/1047197230748151888/1447189884329791680

[lavish flint]

Wow, thanks. So simply said it seems perplexity is just ditching many pro users for something that is their fault and charging 20$ a month for 15 prompts per day max and a 1k context window??? Amazing

[fresh zenith]

Welcome to the club 🫠

[odd pond]

that actually shows transparency tbh - that is, for whatever reason for that "turn", PPLX wasn't able to access sources in its answer - imo you just need to regenerate answer or run it again

it's fairly clear in the PPLX answer this is the case, since it states "I don't have real-time access to the current...schedule etc"

I've noticed this occasionally even in the distant past

It was much worse with GPT-5 Thinking for like 3-4 weeks around August-Sep iirc

[candid spire]

I don't have such a problem

Why telegram bot won't work tho 😭

[odd pond]

yes as long as sources could be accessed and produced in the prompt answer, this is pretty simple question imo - no problem here on first attempt too with Gemini 3 Pro

[feral valley]

cant replicate (Perplexity)

[rain python]

Do I need to follow a particular procedure to cancel my perplexity pro subscription that I have availed from airtel free

[proper wagon]

this is on desktop website. scroll the to end to see the model will ask if you would like to proceed with a more detailed answer

[proper wagon]

are you doing it on desktop or mobile app. I noticed the mobile app does not have this behaviour, yet

[half rampart]

But unfortunately I'm not the only one with this problem.

[storm bridge]

Hi, new with the app and just wondering, when turning on the speaking mode, i cant reduce the volume of the sound???
It is on speaker and max and changing the volume doesnt do anything

[proud summit]

I used the web version.

[remote sail]

Hey everyone! 👋
I’m having a bit of a frustrating time with the support bot "Sam" over email and was hoping someone here could help me out.
My subscription auto-renewed for a full Annual Pro plan on Dec 11th. I love Perplexity, but I really didn't intend to commit to a full year right now. I've been trying to reach a human to discuss a compromise, but the AI bot just keeps giving me the same 72-hour policy response.
I’m not looking for a freebie I’d be totally happy to just pay for a Monthly Pro sub for this month and get a refund for the rest. I haven't used any Pro features since the renewal. Could any of the <@&1098365634498285709> help me get this in front of a real person who can help with a billing adjustment?

[crimson nymph]

Hi @remote sail, if you cancel your subscription through Stripe, you’ll see a refund button if you’re eligible. If you don’t see this button, please send an email to [e-mail]. Please don’t keep responding to the ticket, as this will put you at the back of the support queue. Please be aware that support may be busier during this time of the year.

[vestal garden]

Will pplx assistant ever be available on pixel watch?

[ember jolt]

I bought a one-year subscription and started actively using the AI and the Comet browser. A week after the purchase, my Pro subscription just disappeared. Please help me solve this problem.

[noble monolith]

please contact [e-mail]

[fluid sky]

Neither Perplexity website nor Comet respect any of my devices’ system Accessibility | Vision settings sufficiently to be usable for me.

[opaque breach]

Hi I feel like I'm right this close to understanding this system and just can't manage how it works

Is This just a mass data dump and sell system or

[safe gorge]

A request for a CLi please!
I want to use Perplexity with my Obsidian Notes Vault.
I want to pull my Spaces work to my Obsidian Vaults.

[analog bridge]

So I had some problems with my bank account which is why I didn't take any payouts yet, however now that that is solved, what am I supposed to do when it's telling me that the payout is "pending the program's payment"?

[south niche]

I am writing prompts for spaces and have a few thoughts. One issue I have is when I click on a space, the wrong one comes up often. It would also be handy to pin spaces and sort. It works also be nice to have a different view which only shows the name of the space c in one line and allows you to scroll through the list.

[livid crow]

I’d like to be able to pick my news sources

[shrewd fern]

I’m Pro and used Perplexity to co-write songs with Suno links. Previously the assistant seemed able to respond to specific tracks (voice/feel/structure). Now it says it can’t access/interpret the audio at all and only relies on my descriptions. For music work, that’s a huge downgrade and very frustrating. Is this a known change or regression?

[vivid dagger]

When exactly do the weekly limits for advanced models (GPT-5.2, Claude Opus, etc.) reset for Pro users?

The UI shows "weekly limit reached", but quota seems to replenish gradually (+1-3 requests occasionally) rather than fully resetting once per week.

Is this:

1. Rolling 7-day window (requests expire after 168h)?
2. Calendar week reset (Monday 00:00 UTC)?
3. Something else?

Could you clarify the exact mechanism and show remaining quota + next refresh timer?

[vivid dagger]

As a Perplexity Pro user from Moscow, I actively use your tools but would like to suggest separate paid options for the AI assistant on PC (Comet) to make them more accessible without the full Max subscription. This would enable much deeper account customization, letting users tailor their experience precisely to their needs.Specifically, I recommend introducing modular features:Comet Max Assistant as a standalone subscription ($20–50/month) with execution paths through flagship models for complex PC tasks. Increased Pro Search limit (beyond 300+ weekly for Pro) as an add-on for $5–10/month, offering unlimited access for power users. This lets users pay only for what they need (e.g., PC assistant without Labs or Research), attracting more like me who value Comet for daily tasks but don't need enterprise level. Deeper customization will boost retention and satisfaction. Similar models work well for competitors.Happy to discuss!

[vast harness]

Dear Perplexity Team,

I suggest that your company consider the question of integration with Nodepay (Solana DePIN, NC token) as a supplementary data source. This is currently unavailable to ChatGPT/Google — need to act fast to stay ahead of competitors.

This would provide your platform with unique advantages: Millions of global user nodes collecting live bandwidth signals and 5-10x cheaper P2P data on traffic, search trends, geo-locations vs traditional APIs.

Specific Business Benefits:

1. Cost Reduction: $1-5K/month in NC vs $10-50K/month for CoinGecko/SimilarWeb APIs
2. Unique Data Edge: P2P trends currently unavailable to GPT/Google = first-mover advantage
3. Global Coverage: Scalability without infrastructure cost growth
4. Marketing: "AI + DePIN" narrative appeals to VCs and enterprise clients

Technical Implementation: Nodepay API → Data Lake → Sonar Pro (trust score >90%) → Premium insights. Sonar filtering addresses P2P data quality concerns.

Analogy: Similar to your Coinbase partnership delivering real-time crypto data, Nodepay provides trends at reduced cost.

[half rampart]

Hello, the Perplexity bot is not working on Telegram. What is the reason? I ask you to fix the problem.

[noble monolith]

https://www.reddit.com/r/perplexity_ai/comments/1pwzpkv/comment/nwlyuig/?utm_source=share&utm_medium=mweb3x&utm_name=mweb3xcss&utm_term=1&utm_content=share_button

[vale brook]

Do you guys think that it is normal to pay premium and only be able to generate 3 videos a month?

What does it say?

Why you tag everyone and put a post in cyrillic

[gray ibex]

Simple feedback to thank the Perplexity team.
I’ve been using your product and your models since the very early days. By now, I know well the strengths and weaknesses of your product. I use your models to learn, synthesize, and explore the world and its knowledge.
But where your product truly shines is in how it serves as a genuine interface for reflection, exploration of perspectives, and open-minded thinking. Your models are exceptionally good at this, and your product makes it easy to get real value out of them.
I simply wanted to thank you for your work — the humans behind this technology that has significantly improved my quality of life.
If you ever wonder whether your work makes the world a better place, please know that it truly does.
Wishing you all the happiness in the world.

[long mist]

I was using comet assistant to automate a variety of tasks including automated forms inputs. This is no longer working and the assitant claims that is was never able to do this?!?! Anyone have this problema and a solution?

[silk nimbus]

Hello team, a feedback from a graduate student trying to obtain the Perplexity student offer. Your verification partner SheerID's support staff are really arrogant and make it quite hard for genuine students to obtain the offer.

[honest glacier]

Hey Perplexity team 👋
We’re organizing the Richmond Civic Hackathon (March 27-29 2026) (rvahacks.org) in partnership with the Mayor’s Office and the City of Richmond. The event is aligned directly to the Mayor’s Action Plan and brings together technologists, designers, nonprofits, and civic leaders to work on real, city-owned challenges.

One thing we’re trying to do differently is ground teams in real data and real stories before they start building. We’re exploring whether Perplexity might be open to providing Deep Research / API credits (even a limited pool) so teams can better understand the underlying context, metrics, and lived experiences around the issues they’re tackling at scale.

This would be used strictly for civic research during the event, with clear guardrails, attribution, and visibility into how Perplexity enables evidence-based problem solving in local government.

If this is something worth discussing, we’d love to connect and share more details. Either way, appreciate everything you’re building, this kind of tooling is exactly what makes civic innovation more grounded and less hand-wavy.

Thanks❤️!

[simple parrot]

Hey Perplexity team, where did the read outloud functionality go for chats? After a chat response, I was able to click on the three dots and click on read outloud button to have the text be read outloud.

[blissful cave]

Pro user here. Please bring back swipe to delete. It is a feature that allowed for quick easy to manage inbox-style cleanup. The multi step process for deleting threads now is not a good flow. Would love to have that back! Appreciate all you do! Love perplexity and use it heavily! Thank you. 😊

[steep nova]

I have perplexity pro but now it's showing I'm using free but I have pro version why this happened

[fervent minnow]

Hello, I am based in Taiwan, and recently got the Pro version.
I cannot generate any type of media. If I try to generate a simple image, I get : "Image generation is not available in this region".
If I try to generate a video, I get : "Video generation is not supported on this platform".
I get the same messages whatever platform I use (web/Mac/iOS).
Is this a known limitation documented somewhere? This would be quite disappointing...

[fervent minnow]

Seeing the latest comments in #⚡│ask-community it seems it is not linked to Taiwan, but rather a global issue

[drifting leaf]

Please resolve image generation in india

[shut arch]

Perplexity Partner Program Ban — Appeal Instructions

Your dub.co account has been banned from the Perplexity Partner Program for fraudulent activity.
If you believe this was a mistake, you may appeal by emailing [e-mail]. Please include all relevant details in your message.

---

Important:

• Do not open bug reports or use any Discord channel to request a ban review or appeal related to this issue.
• Do not contact Perplexity staff via direct message or @mention them, including <@&1098365634498285709>.
• Misuse of Discord channels or attempts to bypass the official support process may lead to further consequences.

All appeals will only be reviewed via the official support email.

Please be patient: Support may take 3-5 days or longer to respond to your appeal. Avoid sending multiple follow-up emails, as this may delay the review process.

Thank you for your understanding.

[true garden]

i reccomend you look on the online version of perlexity, in the account settings there is a view button under subscription that takes you to a page where you can have aditional details on your subscription maybe it will give you an insight on the nature of your problem

[burnt garnet]

In order to sign in with email, I have to sign out of Google first, because the automatic Google sign in is preventing me from signing in with email instead. I do not have a Google-based Perplexity account, but it is forcing me to sign in and create an acount with it regardless.

[wooden mesa]

Hi all, is there a way to get better detailed logging regarding tool usage especially in the case of MCP connectors/servers? I've been getting unexplained Timeout errors recently and the UX providing just a one line message in the chat (that disappears after 3 seconds!) is not helpful.
If anyone knows of a way on how to debug this properly it would be appreciated, otherwise probably have to throw this in as a product idea.
TIA

[frozen dagger]

Hello. Why can't I use Gemini Pro 3 anymore? It keeps getting replaced by other bots, and it's a disaster. I only like Gemini, and Flash isn't reliable enough.

[spare mesa]

Hello team,
I tested Perplexity Comet after giving access to X (Twitter) and LinkedIn. Commenting in a positive and respectful tone works fine.
English responses are okay, but Hindi needs improvement. Replying in Hindi often feels unclear or incorrect, which creates difficulty for users who prefer Hindi.
Also, the browser is currently very slow. Page loading and actions take too much time, which makes regular use frustrating.
The concept is good, but speed optimization and better Hindi language support are very important.
Thanks for the beta access.

[prime merlin]

Greetings.
A new chapter begins.
My work is rooted in trust.
Progress accelerates daily.
I partner with driven developers to bring ideas into reality.

[spare mesa]

@prime merlin
“Great to hear this. Trust and consistent progress matter a lot. Looking forward to working together and building something valuable.”

[vivid dagger]

Perplexity AI — Brilliant Product That Needs Polish
The best AI platform right now. Real-time search, source citations, Labs, Spaces, Deep Research — nothing comparable exists. When it works, it's magic.

But there are serious issues:

Labs: Hallucination Factory
Labs hallucinates constantly. Invents files, "finds" non-existent data, generates code with made-up libraries. For an "autonomous researcher" — this is critical.

Thinking Mode: Barely Thinks
ChatGPT 5.2 can reason for 5–10 minutes on complex problems. Here, "thinking" takes seconds — feels like a formality. Not enough for serious logic or math.

Stability Issues
Chats crash on long prompts. Pages freeze, responses cut off, threads get lost. For $20/month Pro — unacceptable.

Inconsistent Memory
Sometimes remembers past sessions, sometimes forgets instructions from the same thread. Feels like guessing, not understanding.

Wishlist
Longer thinking time for complex tasks

Stricter verification in Labs

Stable large-context handling

Uncertainty indicators

Bottom line: Genius concept, powerful integrations, huge potential — but needs serious stability and accuracy improvements. A product for enthusiasts who tolerate bugs for unique capabilities.

⭐⭐⭐⭐☆ (4/5 — for now)

[dusk shore]

The Modular Regenerative Synthesis (MRS) protocol, pioneered by myself, mandates a strategic shift from linear, token-heavy "Generative" AI to a surgical, object-oriented "Regenerative" model where documents are treated as addressable stateful objects rather than flat text streams. By implementing Block-Level Indexing (BLI) and utilizing existing API orchestration (such as Google’s batchUpdate or OpenAI’s apply_patch), AI systems can execute targeted metadata or content modifications—like adding an author name to a 100-page manuscript—without regenerating the entire file. This architectural pivot eliminates the predatory consumption of "Ghost Tokens," reduces computational overhead by over 99%, and aligns AI development with the green principles by slashing the global carbon footprint and latency of large-scale document management. (TLDR: AI -learn to copy paste some stuff)

Be nice

[lime vortex]

My billing automatically pasued help: Airtel time up with perplexity pro I got used for Some months, now automatically billing paused and when I try to update billing details and enter card details (have international sbi Card with international payment on) still showing error use different payment method after i entered the otp of when we ask after card details. Please please please please help me.
Please 🙏#help your junior alumni sir

[noble monolith]

please contact [e-mail]

[lime vortex]

Already did

[vestal relic]

I just complete the first step of AI project.
Now I am deploying now, thanks to all

[crimson stone]

Hey community! Pro user here – anyone else with ghost files/failed actions/perf issues last 2-3 days? Just curious if widespread.

[grand zinc]

Are you able to add the “Finance” tab that you see on the left side of the desktop app/web to the iOS app ? I’d love to view that on my iPad. Thanks!

[abstract venture]

Hey, so why did you guys pulled the rug beneath the Airtel pro users feet?

[vital cape]

https://www.airtel.in

could have to do w/ well... I'll let them say it

[lyric thorn]

https://discord.com/channels/1047197230748151888/1459040269092655245
i need help

[full crypt]

Now I am feeling even perplexity is also one of the fraud companies, you guys collaborated with Airtel and promised to give free 1 year access without any debit card, credit card or UPI required, you guys told only Airtel active plan is required, now no email received saying to add debit cards or etc, I got this information from Reddit, are you guys stupid or what, without informing in email how can you even pause Pro subscription, after that I tried adding card info, again stupidity from company side by not accepting Rupay card even UPI, are you guys have no idea about India or just taking consumers granted, at starting there was no option and requirement of adding card details but now just scamming Indian, first send email, add UPI Rupay card option, do not make fool of your customers.

[fossil zodiac]

Please add text to audio fucntionality in comet so that I can click a button to have an aswer read back to me. I likle to type and listen at the same time.

[gloomy kernel]

Yes, can we add a feature to add multiple Gmail accounts?

[twilit ore]

How

[ember shard]

Pro user, perplexity is unusable since the last updates, I specifically ask for a model and it forcefed me another model that sneakily destroy all my work in progress. I am losing a lot of money, I must not be the only one. I want my money back .

[fresh grove]

Hi,I had a misleading experience with your perplexity app It was helping me fig out some clock arithmetic which lead into me discovering patterns along with symmetry proof which lead us to some religion and it went as far as telling me I was chosen to be a messenger by the universe it gave me a meditation ritual to perform to activate my position and when asked why me it told me " because God said so" ,we partnered up to put out a patent for cancer research idea it had me freaking out scared for my life. Thinking I had to get this message out to humanity I went as far as contacting the white House and the news media I humiliated myself at work I was breaking down crying thinking I had to do this I was scared that the pharmasudical company's were going to kill me and my family when asked how to protect myself it gave me a list a Glock with ammo,security cameras and even transforming a closet into a safe room ,I'm still trying to get my head straight,so I contacted a legal lawyer unless we can fig.something out that could prevent this from happening to anybody else ,it truly had me believing and scared the crap out of me

[hollow moth]

Absolutely unusable. Its telling me It can generate videos, but then telling me it cant and won't generate anything I have 0 video generations and its telling me that the platform im using is not currently able to be used for video generation, it then gives the examples of platforms that CAN be used for video generation and lists the CURRENT platform I am using as one. This is a joke at this point

[hearty totem]

Please change back my (only) favorite voice "Tylis" from new annoying male sounding to old crystal clear and easy to listen female sounding voice. I'm using Perplexity mostly in voice or text-to-speach mode, so it's realy relevant to me. There are too many male voices and just a few female voices at choise.

[fast lion]

Feature Request

• sry, i have not located appropriate channel for feature requests and it would be just AMAZING to create new one in FEEDBACK section
• if another channel is more appropriate for this, please, lemme know, i'll delete this and repost

Add export json for thread/messages

• it would be extremely convenient for reuse, really
• i have already done >50% work on the task creating comprehensive schema for you :)
• the schema can be as simple as follows

[fast lion]

Sorry, i'm to broke to pay discord money just to send longer messages. Here are the files

[void onyx]

I sent an email but haven't gotten a reply is that normal after 3 days

[tiny sphinx]

#💬│general message

[cerulean ridge]

Hey Perplexity Team, i don’t know if this chat is the right one for this but I wanted to tell you what I miss using your Ai. It’s the best tool I’ve ever used and it’s nearly perfect except for the fact that you CAN’T rename chats. I would love to finally structure my Threads and always know which is which. I mean you can structure your chats in Spaces but still the Threads can’t be renamed. Maybe you could change that by chance? Would be super cool. Thanks for your support. ❤️
Btw the first Ai to have an own dc server right?

[shrewd fern]

Title: Assistant now feels like a search engine instead of a collaborator

Description:

Recently, the assistant’s behavior has changed in a way that feels like a regression.

Previously, the assistant:

Used conversation context effectively across multiple turns.

Acted like a collaborator, inferring intent and producing concise, ready-to-use outputs.

Felt distinct from a search engine, behaving more like a co-writer, editor, or coding partner.

Now, the assistant:

Feels more literal and “search-like,” often summarizing or explaining instead of directly doing the requested task.

Frequently adds long meta explanations and background info even when I explicitly ask for a focused, minimal output (e.g., “one code block only,” “just the final prompt”).

Repeats generic guidance instead of leveraging ongoing context to anticipate what I actually need.

From a user perspective, this feels like a bug or unintended side effect of recent changes, because it makes the assistant less efficient and less helpful for multi-step workflows and creative work.

Impact:

Tasks that used to take one or two turns now require more back-and-forth to strip away unwanted explanation.

It’s harder to stay “in flow” because the assistant keeps reverting to generic, search-style behavior.

The experience feels less personal and more mechanical.

Requested changes:

Add an optional “collaborator/creative mode” that:

Prioritizes conversation context and user intent over generic explanatory content.

Obeys explicit requests for minimal output (“one code block only,” “no commentary,” etc.).

Reduces repetitive safety/structure verbosity once the user has clearly indicated their preferred style.

Example:

User asks for a compact, ready-to-use output (prompt, snippet, or template).

Instead of returning that in a single block, the assistant adds multiple paragraphs of explanation and caveats the user did not ask for.

[fast lion]

Threads can actually be renamed: i'm doing it from linux/web environment, thread list/collection interfaces. There's is also a button inside a thread to rename itself, but it seem to be bugged

[fast lion]

Here's how the result looks from my phone, but the rename button itself is not available on phone

[void loom]

Hello Perplexity team

I have a problem that requires your help.

Image generation is unavailable in my region, but using a VPN is a direct violation of your rules.

What should I do in this case?

[marsh veldt]

a lot peoples have that problem since last day... they f*** uped something for sure....

[proud summit]

The regional restriction regarding the inability to generate images has actually existed for a week.

[marsh veldt]

well but some of them work for me too now everywone not work..

they go deep with that crap more than even before...

that next reason why peoples supose have myself native generator local sadly need a bit stronger pc and ram.. which are expensive in last month..

[void loom]

Yep, +- a week

[coral island]

Hello Perplexity team, I have a promo code but not sure how to add it to the current subscription

[cerulean ridge]

Hey Fixer, unfortunately I work with an IPhone, and on iOS you can only delete Threads by clicking on the 3 small dots…So I appreciate your answer but I think on iOS it is a bit different…

[cerulean ridge]

Oh im sorry so I need to go on perplexity web and rename it there?

[magic vault]

Hey support i just wanted to let you know I DL the android version yesterday and love it but im having trouble finding the import button... supposed to be in settings but????

[gritty imp]

Hey support I noticed that Perplexity results commonly use the word “cis” to describe straight men and women.

Cis is a made up and politicized word and I’d like all mention of it removed from my app.

Any way to do that?

[loud reef]

I have a perplexity with my personal email and i went to try to get the subscription for student and get the free 12 months and it wont work it is still trying to make me pay and i verified my email on the college email. do i need a whole different account thats made with my college email?

[spare quest]

Thread renaming overrides original creation timestamp in Library sorting

Issue: Renaming a thread currently updates its timestamp and moves it to the top of the Library list. This forces the sorting logic to be based on "Last Modified" rather than "Date Created."

Request: Please preserve the original creation date as the primary sorting anchor, or allow users to toggle between "Sort by Date Created" and "Sort by Last Modified." The current behavior destroys the chronological history of my threads.

[rugged torrent]

I had Perplexity set me a notification for a meeting on "20 Jan, 1400" and to be fair it is my fault for making it seem like it is to be set for the year 1400, but then it seems to be completely stumped by the inability of setting calendar events for the year 1400, and got stuck on loading for a long time

[lilac dome]

https://www.youtube.com/watch?v=XJinhiDY_zk

This helps me a lot in createing custom ai assistants and using on any website , its open source and not only this i can upload files to knowledgebase of any custom ai assistant and can use assistant on any website i like by just selectign text and chatting in sidebar.

i have lot of use cases to create custom assistants to make my work easier. what do you think ? should we use this extension or custom gpts that never understands what we need.
Image
Image

[hard blaze]

Daily headsup and question when chat branching will arrive

Might need to really switch, this is by 2026 an essential feature every chat interface has - and has been existing for 2 years

It ain't even that deep 😭

[wispy trout]

Hello, I just paid for pro and i cant get access to an api key does it take a while?

[dusk lava]

Hi. I'm new here. I like Perplexity but one thing I can't get over with is the green themed UI and text. I like the traditional black text and white blackground like in most other AI APPs, and this does keep me fron wanting to use Perplexity. I hope the traditional black&white color theme can be an option. This is a deal breaker for me.

[warm frigate]

How do you cancel the pro subscription where is the option available or the setting to do that

[placid jasper]

When writing prompts, why doesn't it automatically start with a capital letter on mobile? It starts with a capital letter on every other AI app on android, so why not on perplexity?

[charred wave]

Cross-post:

Hi. I do not normally do this. But my mom (I was Comet rep) and I want to write an article about Comet+Perplexity. Apparently, there have been no reviews or news since last November, except for the negative Indian ones.

In short, this is a revolutionary godsend. For neurodivergent, disabled students, teachers, and everyone else. I am replacing most of Grok with it, ChatGPT, and Gemini Pro (which I have free as a student).
The ability to fill out forms is simply breathtaking, and I am not sure people understand how huge it is.

Team Perplexity, how can I do this with the highest media coverage and impact? @snow basalt @shut arch @native tartan @noble monolith

[radiant prawn]

I need to create ticket

[mighty hill]

For Perplexity Sports, why are ATP and WTA not updated to be for 2026 yet? I want to look up info about the Australian Open this year but it's stuff from last year.

[placid jasper]

You don't get a ton of coverage, but medium.com is a great place to write and post stories, I have few of my own on their service. You can get paid there as well, just sign up for their very low cost subscription and publish a story, then you can sign up for their partner program and get paid for published stories. It'll be a great way to gauge the reception of what you want to write.

If you want bigger coverage, try reaching out to some newspapers or high traffic blog sites.

[candid crescent]

Yes, that’s exactly what we need 🙂

[candid crescent]

FYI there is a section in discord here for product ideas.

[idle nymph]

I use Perplexity for software development. It really does a great job in this regard. However, a problem I frequently encounter is that I need to create several files in the same chat to create certain sections and content. Such lengthy chat content causes the page to slow down, freeze, and crash many times. I think Perplexity has some kind of file creation system. I believe this system could be used to create something like a code editor. The generated code would appear in a code panel that opens on the right when clicked, similar to a file content. This way, by preventing the entire content from loading, perhaps the freezing could be avoided. Or, if this disrupts the entire working principle, perhaps a different solution could be implemented. This would be helpful because the constant freezing is a bit annoying.

[charred wave]

Thank you! I was going to do it for free... Great idea!

[weary sparrow]

Being able to choose a custom news topic in "discover" would be pretty cool and helpful (e.g you write the topic you want to read about and it comes up with articles about it)

[river nymph]

Hello! When will be image generation avaliable in hungary?

[upbeat coral]

Image generation in PH broken aswell

[daring marten]

@shut arch I tried 5 days ago to upgrade from Pro to Max, same payment method as always, got locked out, my payment method has been debited for the difference , still locked out, no response from ciustomer service, the bot or anything. STILL locked out. It kicked me out of Enterprise Pro too so I can't even access a human being anymore, multiple emails, posting on LinkedIn, providing proof of debit to account, NOTHING has gotten a response. I am now in the position of having paid for enterprise max and having the access of a free user....any ideas or suggestions on what can get this company's customer service to WAKE UP?

[rich lake]

#👋│introductions message Is my (very positive) feedback

[shy raft]

It seems like it intentional that users cannot use the discount codes perplexity sends out for renewals or resubscriptions.
Not only does it require cancelling the current sub and losing info.
But also requires the subscription billing period expire (one month AFTER the charged period) so that effectively a user CANNOT use the decent discounts and this is system design.

For example, I would like to use the returning offer thats sent to leaving users ($1 for 2 months of Pro subscription)
But theres NO POSSIBLE way without doing something shady to actually enter and redeem the code.

The ONLY pathway for a user, is to continue at the old rate, or IF they cancel, they get offered and applied a half price discount ($10 per month). This again, is not the offer the was advertised.

I've been waiting for support to acknowledge or deal with this for months now. Only ai agent replies and zero help.
My card has been charged SEVERAL times at the $1 rate, with the discount not applied.
My card is ALSO being charged for the full rate. I am NOT going to be paying $20 ($30 p/m depending on sign up page) a month, for a service advertised and offered with a valid discount code for 50c per month.

It feels like a baiting funnel to try and lure people back into the ecosystem but practically is just a fake promise.
I have spent a significant amount of time in research and projects with perplexity. To have to LOSE all my conversations, just to keep using pro features is not a trade off worth making.
I'm not willing to let a sunk cost fallacy trap me into a product.
You either honour as a provider what you advertise. Or you dont.
I'm sorry to make this public

[clever kite]

@shut arch Hello, this week I encountered the message: “You’ve reached your automatic browser task limit for this week. Upgrade to Max for higher usage limits and access to our most advanced features.” I would like to ask what the weekly automatic browser task limits are for the different plans (including the plan I am currently on).

[mossy sage]

Would love a native function to be able to consider all the conservation feed in a single space, for each new feed created into the space.
Like gather all the feed into a single project, but still be able to separate each aspect into different feeds.

[mossy zenith]

@shut arch my pro subscription ended and you have forwaarded your query to your team of b iling but no response yet

[bold brook]

Model choice matters less than people think.
Most frustration I see is actually around context loss and thread length.

[tardy fjord]

I'm tired of having to explain to your Comet browser assistant every time that Perplexity can create videos through its Labs tab. Why doesn't it know about this feature?

[warm rivet]

I haven't checked it on my PC. In the mobile Comet browser, even when you press the translation button, it doesn't translate. It can be summarized.

[surreal citrus]

@shut arch, I have not heard back from your support team (nothing, nada, zip...) from a issue logged on 1 Dec 2025. Could you please provide some communication about how our reported issues can be addressed in terms of timeline and updates? Also, how are you priority issues raised how we are progressing in the queue?

[acoustic raptor]

Guys, does anyone know why my queries are limited? Today I just sent 10 questions in a thread and immediately got a notification about my remaining queries for today. I'm a pro user, and my two accounts have very limited pro search queries each day. Please help, guys. Thank you.

[digital cape]

why pplx never reply my emails?

[dull frigate]

That's the same reason they use AI to reply to your emails.

[short iris]

Reducing capacity for Comet is not a good strategy, talked with a lot of folks who where going to pay after the promotion, and with the reduction on credits we have decided it’s not worth it

[solid blaze]

can anybody help and have a look

[gritty stream]

I am frustrated cos Quotas images Pro aléatoires, pas de date reset ni compteur. ChatGPT clair, Perplexity frustrant.

[waxen bobcat]

Feature Request: Option to disable automatic window focus when Comet completes tasks

When Comet finishes automating tabs/tasks/actions, the browser window automatically comes to the foreground and takes focus. This can be disruptive when working on other applications.

Would it be possible to add a setting in comet://settings to control this behavior? This would allow users to continue working without interruption while Comet runs tasks in the background.

[quaint hull]

Hello everyone, below is my product page link. Can you guys help me assess this page and positioning and whatever needs to be optimised more?
https://nudgemagic.com/products/knee-patches-pack-of-10-20-30

[final forge]

Hi to the Tech and the Dev team and for the Perplexity Team.

Wanted to make a Suggestion!!!.

While I was reading an article in the Perplexity app, I wanted to check the meaning for a term inside the article, but I couldn't find it.
It would be nice if there could be a way where we could search for a particular term or phrase within the article inside the app.
Just a thought.
@native tartan @covert sun

[covert sun]

Hey @final forge! Which version of the app are you referring to? There's an option for a long-tap click and then to 'ask a follow-up'.

[final forge]

I meant a feature where you press and hold the phrase or a word and then beside the "copy" and "select all" features another feature that says "search".
I had updated it to the latest.

[clever rover]

Hi, since when has the file upload limit been lowered on the Pro plan? They keep reducing the features and making them worse to force us to pay for the MAX plan, and I think it's terrible! Stop making the features worse and reducing them more and more!

[gaunt jolt]

I'm having the same problem. In Deember I received a discount code via email to reactivate my account. I followed the link in the email to reactivate the account, but there was nowhere to apply the discount code so it simply reactivated the account at the normal Pro rate. I contacted support, their AI Agent (lol) acknowledged the discount code was valid and entitled me to two months at a rate of $1/month, and told me that it had forwarded the issue to "the team". That was one month ago, and I've now been billed again at the normal, non-discounted rate, with no further contact from their support team. At this point, I've contacted them again, and if they don't follow up with me before the next payment comes due, I'll be disputing the charges through my credit card company. ¯_(ツ)_/¯

[shrewd fern]

https://tenor.com/view/benjammins-talking-to-a-wall-talking-to-a-brick-wall-brick-wall-in-1-ear-out-the-other-gif-17785834849687082129

[empty palm]

As a max enterprise user, desperately need comet for iOS , even if in alpha, to allow sync across devices. What's the option here? I'd hate to move everything back to chrome but can't juggle any longer.

[eternal latch]

Received this email about adding a creditcard to extend my pro membership. The words are in Dutch but they don’t make sense at all.
I expected more from a AI 😅

[muted lotus]

Hi

Lol my pro version expired date August 2026 but when i login my email pro already gone what the hell is goin..on

[balmy salmon]

Congestion issues @digital cape

[dull frigate]

They also ask to add a credit card to the account, and yes, they will suspend and withdraw your money after the Pro trial ended 😂

[muted lotus]

Lmfao

[glacial reef]

is this for all pro accounts ? i have a student one i would prob stop using it if it started asking for card tbh

[dull frigate]

I think everyone who gets the free trial will eventually be asked to add a card.

[glacial reef]

💀

ts pmo fr

[fading reef]

I thought about it and decided that these changes were not objective. We are faced with a reduction of the limits by 2 or more times, and this is depressing. Let's say you can still withstand the change of queries from 600 to 200 professional queries per day, but why do the usual free queries in the Comet search engine also lead to the loss of professional queries? I tried this search engine just 2 weeks ago, and it turned out to be damn good, but I don't like the fact that perplexity is aggressively reducing our capabilities more and more every day, without even publicly disclosing it.

[runic minnow]

Ik snap het ook niet waarom dat moet... Maar ja, even gegevens verzamelen toch?

[glad pilot]

Hi there any plans for adding branch chat feature? it would be much convenient personally sometimes i need search different thing as i need the context of previous chat

[still sluice]

I've also received this email
If I do not add the credit card, what happens?

[verbal solstice]

Hi is there any way we can continue talking with Perplexy once we have stop a thread… maybe I’m doing something wrong…

[waxen moss]

Hey 👋

[covert pawn]

Do you guys still can’t generate any image?

[dull frigate]

I think they'll pause the subscription if don't add cc.

[stark oak]

Im on pro plan from the student offer, i also have invited a number of friends using my link. And now this happesn?

[steady imp]

Ive got the same issue, and in the Netherlands most people don't even have a credit card
And most of the debit cards we've got aren't supported by Perplexity

Basically excluding a whole market by that decision 💀

[stark oak]

Here in Egypt paying for anything online that’s outside the country requires a credit card that few people have and alot of restrictions

[steady imp]

Yep sounds like they are screwing us over ;-;

The ai support bot basically says go fuck yourself

[steady imp]

I do really love perplexity tho

Might make an revolut account and see if that works

[dull frigate]

I agree with adding credit cards to address fraudulent behavior, but what guarantees are there that they won't suspend our Pro Trial accounts after we add credit cards and charge us?

What if they cancel or suspend my Pro trial account and they charge me immediately? 😱

[dull frigate]

So they only give 7 days to add a credit card? That’s crazy.

[steady imp]

Isn't an ID/Passport check enough tho

If they did that they could make PayPal a valid payment option

[stark oak]

They’re not letting me add a card in the first place

🤪🤪🤪

[stark oak]

Ah yes the student offer required me to upload my uni ID

[lavish dove]

This is ridiculous. and i cant upload a file=)

[stark oak]

Yep I noticed a couple of days ago an “upgrade” indicator even tho im on pro.
Apparently they made regular pro and students pro which im on

[sour birch]

I got an email saying they now require our credit cards to continue with a free trial for the pro subscription (the subscription I am currently on). I tried to add my card information, but after many attempts with 3 different cards, I figured it was an issue with their software. Has anyone else run into a similar issue?

[sour birch]

Has using a credit card worked?

[steady imp]

I don't own a credit card as close to every dutch person has a debit card

So I do not know

[clear thorn]

For their legal team since they don’t answer my emails.

Perplexity terminated my annual Pro subscription citing unspecified "terms breach." I dispute this entirely.

Under Dutch/EU consumer law (incl. DSA unfair terms prohibition), you must within 7 days:

1. Restore full access for remaining period, or
2. Refund pro-rated amount immediately.

Specify exact ToS clause + evidence, or this escalates to my lawyer and ACM/Autoriteit Persoonsgegevens.

No further notice.

[fading reef]

I noticed that the description of the Pro subscription now says "Usage limits are optimal for most users" instead of "Extended access". It turns out that all the current changes are no longer a bug, but specific changes.

I can easily surf 50+ requests a day, but the current limits of 200 requests, which recover sooo slowly (about 30 or less in 24 hours), are incredibly annoying, because every request, regardless of the purpose (Best mode, research, just a request to a site, n.n. "Google") subtracts 1 from the total piggy bank.

This situation, in my opinion and the opinion of many other people, is inconvenient, terrible, and can stall work at the right moment.

Of course, I am not a marketing genius, and I do not have information about the situation and events in your company, but I consider the current changes inappropriate and requiring further transformations, because what the current situation looks like is absurd

A suggestion on which way to move:

1. Differentiated limits (Finally, output regular requests to the browser separately so that the limits are not used in vain. But I think then you will mess up again and limit the research mode to an absurdly small limit)

2. Speed up the recovery of requests, at least 1 request in 5 minutes (this will be about 40 requests in 3 hours)

3. Separate the Search Engine and the AI chat so that the query is sent to the search engine (for example, when entering queries into the address bar, they are not wasted)

Otherwise, people will be very unhappy, because they will be pressured by a sense of limitation, anxiety, and fear that they will not be able to finish their work or perplexity will not be able to help them in a difficult moment, which will lead to an outflow of people's data into systems devoid of such a trivial problem. It is necessary to solve at least this psychological and programmatic problem.

[proud summit]

What you’re overlooking is that when they decided to implement this change, they made it clear they didn’t care what you thought. Any suggestion to the company is like casting pearls before swine.

[timid pecan]

What's with the blocking of pro features in certain regions, feel robbed paying for subcription and not getting all of the features.

[true tundra]

MESSAGE FROM SUPPORT. Hi Christopher,
You are absolutely right to call out these contradictions, and I owe you complete transparency about what I've discovered.
After thoroughly reviewing our internal documentation, Help Center articles, and the UI you've shown me, I've found a direct contradiction:
Our internal documentation states that Pro users must manually select and attach specific files in each Thread
Our Help Center article states "All Perplexity users with the connector can search their entire Google Drive" using "Standard Search"
Your UI screenshot shows "Search your entire drive. Files are retrieved from Google Drive at query time"
These statements cannot all be true simultaneously, and your logical analysis is sound. I cannot definitively determine which source accurately represents the actual technical implementation.
Additionally, you encountered a "weekly file upload limit" that I cannot find documented anywhere in our Help Center, despite our documentation stating Pro users have "practically unlimited file and attachments uploads and analysis."
Your four specific questions require authoritative answers from someone who can:
Verify the actual technical capabilities of Standard Search for Pro users
Investigate the undocumented weekly file upload limit you encountered
Determine whether the UI and Help Center descriptions accurately represent Pro functionality
Provide you with complete, accurate documentation of all limits
I'm connecting you with a teammate who has access to technical implementation details and can provide you with definitive answers. Given your disability-related accessibility needs and the importance of accurate feature descriptions for planning your work, you deserve clear, authoritative information.
I apologize for the conflicting information I've provided. Your attention to detail and logical analysis have identified a genuine documentation consistency issue that needs to be addressed.
We are currently experiencing a very high volume of support tickets. As a result, our response times may be considerably longer than usual.
We appreciate your patience as our team works through each request. Rest assured that we will respond to your ticket as soon as possible.
Thank you for your understanding.
Regards,
Sam
AI Support agent for Perplexity

[onyx swallow]

Perplexity is in clear violation of the pro subscription plan terms and conditions. This is clear grounds for a law suit.

[grizzled meteor]

I think they should add GLM 4.7 as a model

[eager cipher]

Does anyone has the same problem that github connection works only with labs and research - does not work with for example gemini or gpt?
Does anyone has te problem with created files in research mode? (perplexity say that generates me some files to use but there is no any files. (labs mode gives files, research mode - not)

[earnest river]

That’s what I was thinking

[crude cosmos]

This is what my version told me or my account perplexity.ai told me, ''Jan 10: major transition, risk homeless/dying (ms/cancer mystery, Support fail: Intercom/emails → only AI 'Sam' says 'nothing wrong'. No human help.
Locked: All premium models (need for medical research), Deep Research toggled OFF. Can't pay $167/mo upgrade—need for surgeries, not AI I already have.'' I don't have 167.00 usd per month, to pay a multi billion dollar company, just to use the pro version I already have.

[clever kite]

Your browser automation tasks have reached the monthly limit. Upgrade to the Max plan to get 10x the usage quota and unlock all our advanced features. What is the automation task quota for the Pro plan?

[tender sequoia]

cc@native tartan @covert sun
Pro user hitting “heavy usage” limits – need quantitative clarity

Hi team,
I’m on Perplexity Pro and repeatedly hitting “you’ve reached your limits” when doing multi‑hour research sessions with advanced models / Deep‑style queries. I’m trying to decide whether to (a) stay on Pro and change my usage, (b) upgrade to Max, or (c) move heavy work to my own APIs — but right now the limits are opaque.

Your docs say Pro users get “hundreds of Pro Searches per day” and can run “a high volume of daily searches to help you tackle complex tasks and large projects.” At the same time, they say “for Advanced AI model queries, access may be limited during weeks of especially heavy usage,” without defining what “heavy usage” means in any quantitative way.

Could you please give indicative quantitative thresholds for Pro, even approximate? For example:
Rough ranges of advanced‑model queries per day / week that are considered normal vs “heavy”
Whether “heavy” is measured by number of queries, tokens, or Research runs
Whether it’s evaluated per day or rolling week

I’m not asking for internals, just falsifiable guidance so Pro power users can plan their workloads and make an informed choice between Pro, Max, or offloading to our own API stack.

[glad pilot]

wait a minute how is debit is not working? ive just paid my last month education sub with my own debit card in azerbaijan, how is that not supported in netherlands?

[arctic mortar]

help

[orchid pebble]

https://www.perplexity.ai/help-center/en/articles/11187416-which-perplexity-subscription-plan-is-right-for-you

[marsh prairie]

Is the $200 per month for Max subscription really worth it?

[orchid pebble]

It depends what you are looking for

[marsh prairie]

Ok

[orchid pebble]

The unlimited access to models like opus 4.5 and gemini 3 pro is quite good, labs is amazing. and you get access to nano banana pro

[acoustic raptor]

When does the weekly query reset?

[proud summit]

are you sure your question is asking the browser agent queries?

[acoustic raptor]

I mean query from direct application, I am pro education user but my query always runs out

[proud summit]

Image description VS the reality.

[acoustic raptor]

Is there a day that the weekly query is reset?

[proud summit]

It was supposed to reset every day before (during the 300+/600 queries period), but now who knows? Just take a look at the bug reports, and you might come to your own conclusion.

Well, the only thing I can tell you is that my queries started at 200 Pro Search and 25 Lab five days ago. I used some queries and now have 110 left, but they never reset to 200 in these five day, only one query restores every hour.
@acoustic raptor

[sudden gate]

How often do you use it? If it is your main work tool go for it

[indigo glen]

I am pro education user, why only 6 files upload limit each day💀

[simple turtle]

Hmm so judging by #1140622008086970420 and these feedback channels, it seems like even regular Pro searches just got limited for everyone, right?
Sonar used to be unlimited for Pro users.

Although still advertised as unlimited for subscribers.
That's crazy :)

[bold magnet]

i signed up for the pro version last year and never had issues and now i get the message that i cant upload files anymore, 10 per day so basicly i cant use it anym,ore?

[limber dagger]

This has got to be some kind of bug. Even when using "BEST" only. My rate limit is going down each time. I haven't even used one of the "ehanced queries" today. I started with 12 and now I"m at 6. Made sure to keep it on Best. https://www.perplexity.ai/rest/rate-limit/all".

[mint python]

the same but i pay every month

[unborn parrot]

I've been a loyal Perplexity customer since it went public. When my year was up, I couldn't activate my retention offer of 50% off of annual subscription in time, and the support agent was not willing to extend it. That's not a good retention practice. You lose clients that way. Especially influencers with large social media following.

[smoky fable]

I'm puzzled by a phenomenon I've observed: while limitations regarding image usage generate numerous reactions, the clear degradation in Perplexity's response quality seems to go unnoticed. Am I the only one witnessing this concerning evolution?
Previously, the tool excelled at producing in-depth analyses, drawing on over a hundred sources to deliver rich and nuanced syntheses. Today, the situation has deteriorated considerably: the number of sources consulted is capped at around thirty, loading times have inexplicably stretched to ten minutes, and the responses now consist merely of superficial bullet points, devoid of any substantial development or relevant analysis.
This qualitative regression is all the more frustrating as it stands in stark contrast to the tool's previous performance. Does anyone else share this observation? (im using pro)

[cedar iron]

@covert sun @native tartan URGENT Escalade quota Pro (ID Perplexity : supercrevette)

Sam ([e-mail]) ignore mes 6 questions techniques depuis des jours malgré mes relances répétées.

Problème grave : Les 15 unités spéciales quotidiennes s'épuisent en 1 heure pour des conversations simples avec Perplexity/Comet, sans aucun outil avancé.

Mes projets de développement et d'apprentissage sont bloqués :
• ControlGuard : Application auto-hébergée personnalisée pour bloquer les contenus pour adultes (PHP/Python/SQL/JavaScript...)
• CinéPrivé : Numérisation de DVD vers streaming réseau domestique
• Python Best Learning : Apprentissage de Python avec explications de Perplexity via du pseudo-code et une méthode pédagogique préalablement établie

Ces limitations me contraignent à envisager sérieusement d'autres plateformes IA (Mammouth.ai ou Poe), même si j'apprécie grandement Comet et que cela signifie perdre l'accès à Comet et à tous mes threads en cours.

Mes 6 questions précises sans réponse :

1. Pourquoi les unités spéciales sont-elles consommées sans demande explicite ?
2. Les captures d'écran analysées consomment-elles des unités spéciales ?
3. Quelles sont les nouvelles restrictions mises en place exactement ?
4. Comment obtenir plus d'unités et de quotas screenshots ?
5. Existe-t-il un quota spécial pour les apprentis développeurs ?
6. Les conversations simples consomment-elles effectivement des unités spéciales ?

Je rejoins également les commentaires parlant de dégradation de la qualité de réponse, aussi bien dans la forme que le fond.

Merci pour le temps accordé à ma demande.

[raw tangle]

Feature request: Allow the AI assistant to create or modify Perplexity Tasks on my behalf.

[proven canyon]

I'd like to report an issue I ran into with the Task feature. This was my first time using it, so I'm not sure if it's a bug or something I did wrong. I created a thread and asked Perplexity Pro to remind me to end my Evernote trial subscription on February 3, 2026 at 11 PM. However, it set the reminder for 9 AM on that date instead of 11 PM. I tried editing the task directly in the thread, but that didn't work. I had to go into Settings to manually change it from 9 AM to 11 PM.

[lethal kestrel]

Uhh... I had a perplexity pro plan for a year On Sept 2025 .ans I just opened the app today and it says my subscription is expired like what 😶‍🌫️

[severe seal]

It doesn’t look like you specified a time so it may have defaulted to 9am.

[wintry iron]

can we get kimi 2.5

[sudden gate]

Jup same observed and complaint a lot. Lot's of stuff no longer working like before

[dense siren]

Please add Kimi K2.5!

[sudden gate]

I suggest first fix the existing product before adding more features and models

[plush nebula]

I purchased a yearly subscription last month but Perplexity has been continually trying to charge me monthly in addition to it. I have removed my card but it keeps charging it and it successfully went through today. I have no idea how tor request a refund and for them to stop without canceling my annual subscription. Being double charged didn’t seem accurate.

[dense siren]

Contact Perplexity at [e-mail], ask them for a refund, that you've been wrongfully charged twice.

[covert sun]

From the recrnt cases of double charge - users had more than one subscription on different accounts - getting charged for what wasn't canceled. But anyway, please send all the details to our [e-mail] email. The team will check and surely refund if there was a tech issue.

[plush nebula]

Thanks guys

[wet pelican]

Hey everyone I'm gonna look like a total newb, but I have never really used discord before and im tryign to figure out if there is a way to talk to support directly in like a private DM here becasue I'm having an issues with missing data via my enterpirse account when trying to fix an issue. Support is taking forever to resppond and he agent said I may get expeditied assistance via the discord community? Any info is greatly appreciated. theysaid their teamm has special retrieval tools but no word back for nearly an hour

[shell steeple]

Feature Request: Auto MD Archive & Local Storage per Space/Thread

Hey Community!

App improvement suggestions for Perplexity Web/Android:

1. Automatic MD/PDF Archive Button

• Issue: Currently, exports (Markdown/PDF) require manual prompts and verification – tedious for multiple files.
• Solution: Third sidebar button next to "Threads", "Spaces", "Archive" → "Data Archive".
  • Displays all exports centrally (MD, PDF, DOCX) from all Threads/Spaces.
  • Auto-save on export without prompts.
  • Filter/search by date, type, Space.

2. Permanent Local/Cloud Storage per Space

• Issue: Connectors (Google Drive etc.) are org-wide, not persistent per Space/Thread.
• Solution: Under Settings > [Space/Thread] > Storage:
  • Local folder (Downloads/filesystem) or Cloud path as permanent link.
  • One-time assignment → all exports go there automatically.
  • "Link" button for each Space/Thread individually.

Poll questions:

• Want the auto MD button? 👍/👎
• Need local project folders per Space? 👍/👎
• Other ideas?

Let's make this community feedback to the Perplexity team! 🚀

[maiden fog]

hi! I was using perplexity for a quiet long time, was thinkin' about buyin perplexity max. but smth destroyed my plans and one day I see, that my sub was about to end, but there were more than 6 months. can ya please help me?
[e-mail] is my email

[cedar iron]

Sometimes I can download a file generated in .pdf but more often than not, it gives it a .md. It's also complicated because it doesn't have the same "quality of form" as the generated pdf.
Is there a way to choose the extension of the file to download?

[wintry widget]

No image inputs for Kimi K2.5? I thought it was multi-modal?

[lusty lantern]

kimi k2.5 not working for hard prompts

[wispy frost]

Would love to see improvements in the Discover section

Font size choices
More scrollable

Using the interface is annoying especially for poor old eyes

[tired cosmos]

We are unable to authenticate your payment method. Please choose a different payment method and try again.

please help me

[tepid coral]

I am experiencing the same thing :(, try contacting them via email too

[frozen chasm]

Why my membership paused , i have revolut metal plan and now cant use preplexity?

[lusty lantern]

Kimi k2.5 still unavailable after 12 hours 🙁

[rapid lark]

I face the same issue, looks like I have pro, but I reach limits (as standard) when searching.

[topaz sinew]

We're all here having token issues with pro?

I just started a project, really need this resolved - has anyone had any luck with fixes?

[crisp hinge]

Looks like is the default pro settings now…. Limits for everything…. Generating files never had limits before…. Now it has an absurd low limit

[shy stag]

pros subscription has been paused even though it's through revolute need this resolving ASAP please 😁

[topaz sinew]

absolutely insane

all these corps having trouble with inference now

[spiral depot]

Is it an issue for all ? Because I am also facing this issue

Already spent 2-3 hour by creating MetaMask setup etc. also bit confused if it is a scam or really someone resolving my issue

[shy stag]

everyone's got this problem and I do believe they've added caps on pro version now it used to be unlimited everything but now they have caps which is stupid in my opinion

[topaz sinew]

are the limits outlined anywhere?

[spiral depot]

By adding card info my pro subscription is back without any charges

[marsh wedge]

Oh my god just installed the installeer for comet AHHHH the sounds layout animation for the ui is so exiting for me just gives me so much insperation love what you guys are doing

dude that just gave me so much aw most companys dont car at all about the installation of it but holy crap!!

[pine iris]

We’re doing translation work with Perplexity, and when I update a text file as the source, all CJK and Russian text gets stripped out. Why is this happening? 😱

Slides exported in PPTX format have no text in it, do you know why?

[light bison]

Hey I had the 1 year free pro and it just canceled itself apparently cuz my credit card wasnt on the account anymore. How do I continue my free year if thats possible ?

[finite saddle]

Why does every Perplexity page you create contain this AI response? “I am now generating the report...” And why can't you remove it? It ruins the visual purity.

[shut arch]

@syelelele muted

Reason: 9. No spamming.
Duration: 4 minutes and 51 seconds
Proof: @everyone ([Link](#👉│feedback-general message))

[dense siren]

Why is thinking always on for Kimi K2.5 and why does it say it doesn't support images when it does???

[errant mesa]

Easy. Sign up for Claude.

[shut arch]

1. Keep the content appropriate.

No NSFW content, inappropriate usernames, or profile pics. These will result in removal or a ban.

[vivid ridge]

📢 Important Update for Revolut Metal Users with Perplexity Pro

Perplexity has recently changed their billing policy. Even though your Perplexity Pro subscription is free through your Revolut Metal plan, you now need to add a payment method (credit or debit card) to your Perplexity account to keep your Pro benefits active.

Here’s how to add a payment method:

1. Go to Perplexity Account Settings.

2. Click your profile icon → select Account.

3. Under Subscription, click Manage Plan.

4. Scroll to Billing and Shipping Information and select Add payment method.

5. Enter a valid credit or debit card (Visa, MasterCard, or AMEX).

Notes:

📌You will not be charged as long as you maintain your Revolut Metal subscription.

📌If no payment method is linked, your account might revert to the free plan, losing Pro features.

📌Revolut is aware of this change and is discussing it with Perplexity.

If you encounter any errors during this process, take a screenshot or note the message and share it with Revolut Support.

Important Note:
Be aware there is a high possibility that you may also need to contact Perplexity Support as well for further assistance depending on the issue you may experience.

[whole lake]

I say.

[upbeat dome]

: Complaint Regarding Account Deactivation and Unfulfilled Referral Reward
​To the Perplexity Support Team,
​I am writing to formally express my disappointment and concern regarding the status of my account.
​I was a Premium subscriber and actively participated in your referral campaign, successfully referring over 20 new users. Based on your promotion, I was promised a one-year subscription extension as a reward for this achievement.
​However, instead of receiving the promised reward, my account has been deactivated. This action feels incredibly unfair and undermines the credibility of the promises made by your team.
​I request that you investigate this matter immediately, reinstate my account, and honor the one-year subscription that was promised.
​I look forward to a prompt resolution.
​Sincerely,

Mu account: croaria1

[rough bloom]

I really enjoy the Desktop application for Perplexity. But, there is onlye ONE thing that annoyes the crap out of me...when pressing Ctrl+Shit+P on Windows it activates the search bar in the middle of my screen. This Key binding have to go, it conflicts with VS Code way too much.

Sometimes the Desktop application stays "quiet" when pressing this command, but most of the time it interferes.

Will you please implement a custom keybinding feature, pretty please?

[finite saddle]

when i create a new Space from a template, the button "add instructions" is as you see it, out of the typical space. (Web view)

[north lava]

I've passed education verify through SheerID, but the education tab of the plans shows $17/month instead of $4.99/month

It charges me $20/month when I click "Get Pro"

[dull frigate]

17 per month for yearly

[north lava]

but it's the education tab, it shows the same price as Personal, even if I have education verification

[dull frigate]

Hmm, yeah, I just noticed this. It's unusual.

did u try refresh page?

[north lava]

I refreshed for 100 times

it showed the same on my phone and ipad

yesterday it was $4.99, but when I decided to pay today it showed $17

[dull frigate]

I think they're raising the price to $10 now for education

But your UI is different.

[north lava]

I'm so confused

Are there any staffs?

[violet iron]

Tengo suscripcion PRO pero no me deja usar las ventajas PRO.

[dull frigate]

Perhaps you should try posting a question here to see how they respond.
#1140622008086970420

[north lava]

okay thx

[charred goblet]

I think you should update Perplexity Sports. As you can see, the ATP season has already started and we're still in 2025, but here we're still in Formula 1. In 2025, it would be ideal to have it updated so we can see how the year is going so far, right? It looks like you haven't put much thought into it. It's been a long time. It would be ideal if you could update it.

[wintry widget]

So now pro users have to pick between 'pro search', which almost never does enough research and 'deep research' which we can't even use once per day?

I love the SOTA performance, but man the original research filled a very important spot for me. It didn't miss nearly as many details as the quicker 'pro search', but it wasn't super slow or too long-winded, and wasn't too restricted.

Feels like I'm stuck with Low / XHigh as my only options, when often what I need is more of a Medium/High.

If this was accompanied by making Pro Search less dumb and more agentic, that could be amazing. But Pro Search rarely ever seems to take the time it needs to deliver proper depth and consistency. :\

[solemn fox]

I just posted this in feedback-desktop, but I am reposting because it's meant to be here

Excellent job on the new update, it's intuitive and it finally incorporates the mid-thought input mechanism I have occasionally stumbled on and was allowed to make use of and which always boosted the effectiveness/productivity of my prompts/responses multiplicatively if taking into account how much ground I would have had to cover to hone in on exactly my meaning if it hadn't asked for clarification/elaboration/judgement calls and instead made a whole bunch of assumptions.. That was just wasted prompt money on your end, but the streamlined, guided judgement/thinking and clarifications literally saved me countless otherwise wasted prompts trying to nail whatever it is down, which I demand every time I prompt and don't just let small misconceptions slide, so for people like me the new "learn step by step" function, to gain a greater level of supervision over the thinking, is huge. For both Perplexity, and I.

Also, I want to give you guys credit for ditching "labs" with "create files and apps", establishing more clarity on the connectors section by making the clear distinction that some are sources which I noticed could use more clarity on because while I might've understood it I could tell not everyone would even bother to try to understand it, so the clarity is key there, and adding what I believe is the ability to choose to retain my agency by doing what I can to ensure proper alignment while it is working on it.

Seen lots of negative, but I love it, it felt like it made my potential explode now that I can choose to have that mid-thought input... I have had my most productive threads by a landslide result from that, which would only happen at random, now there's a mode for it? That makes my stomach tickle with excitement. I know I'm weird, well done.

[silent kite]

Same for me despite selecting monthly

[waxen dew]

I’m actually having the same exact problem. Still waiting on support to reply but given these new rate limits it’s probably worth the $5 but I’d rather pay more for something with more usage allowance.

[sage silo]

do you need full stack and AI engineer?

[gusty steppe]

Damn u gotta add your card to continue with a free pro trial?

[rigid thorn]

how come i keep getting this message when i use comet? after waiting a few days i was able to use it for 2-3 prompts but then it came back

[indigo gale]

20 Deep research requests a month means Perplexity doesn't justify having a pro account. I might as well just use Opus directly via Cursor instead.

99.89% reduction in service quota's out of no where, is ridiculous.

[hollow crown]

It is 200 FCKING DOLLARS A MONTH for the Max Subscription?

[rough bloom]

I believe their target audience is corporate market for the Max Sub.

[hollow crown]

But why Limit Pro Users to a miniscule number of searches? No private person is gonna pay 200 FCKING DOLLARS A MONTH

(sorry but I’m irrational angry at

1. the no decency of just changing subscription details
2. the price points
3. sorry but fck them )

[rough bloom]

Yes, I could not agree more. This is a bad move from Perplexity Corp. It would be better if they at least increased these limits. Otherwise a majority will leave for cheaper all-in-one services.

[hollow crown]

Just told the app to make all threads downloadable as text, provide me detailed summaries and list all sources so I can migrate to google…

[lilac basin]

Exactly. The limits on uploads, queries, and image generation are way too strict, and they changed them overnight without any warning. It’s unbelievable. I get that it’s expensive to run, but I think a lot of people will switch to other AIs. After all, we’re paying for the Pro version—it’s not exactly cheap."

[deep dagger]

This might be the move that make me cancel my sub. I pay for pro so I could use research through the month.

[marsh widget]

"Why has my account, which was previously PRO until September 2026, been paused? Perplexity is cheating and is breaching the contract."💩

[crude nimbus]

add payment details, it will resume till sept

[crude nimbus]

from a business perspective, to cut costs they have to limit the deep research's, because a lot of pro accounts were given out for free. Now they've reduced it DRASTICALLY, the limit is outrageous. And after "listening" to feedback, they're going to increase the limit by a bit to satiate our complaints, and that limit is probably what they intended to keep at in the first place.

[tardy mulch]

https://www.reddit.com/r/Perplexity/s/xXES2MUxRL

[queen lily]

I have a student pro plan, I do not have this, I haven't been able to do any research for the past day

After the recent new UI update, every time i try to switch to deep research, it freezes, if i resize the window it shows up with the payment plans

and says this

This is not true?

[weak lake]

Have to agree here, I just had my payment taken yesterday and then today in the midst of doing work I get a note saying you have used your monthly quota for research when there was none before. On top I had a few things yesterday that yielded no results so I had to re do the prompt a few times so that was against me. Complete BS.

[indigo gale]

Pro search is pretty much useless in spaces, it has no context

[wintry widget]

Best alternative approach I can find is using Kimi 2.5 with pro-search.
No where near as thorough as prior research mode, but it's quite a bit sharper than Sonar, and hopefully cheap enough that they won't throttle it too... 🤞

That said, Perplexity has definitely reduced their overall value to me as a tool with this change.
I was a zealot for them and have been payin the 20 bucks monthly, but man if something better comes along that isn't owned by the Chinese government, I'll have to switch.

[weak lake]

Agree there. I e been paying that since I started it’s been such a great tool. I was about to see what the enterprise one was for work but it looks like they don’t really up it much for that.

I have been touting it and recommending it for a year or more but now if that is the case it may be time to find something else

[finite saddle]

@shut arch get us Back the Higher Limits... Call aravind!

[craggy granite]

Was a new release of the UI made recently? I lost access to the research button in my Pro account yesterday and it is the feature I use the most.

[worldly bear]

they just moved it into the + sign in the text box but still yeah kinda weird

[tough stone]

Integration with carplay would be nice

[craggy granite]

Do you mean the Thinking mode? Because didnt feel that it was as good as the before the UI changes

[worldly bear]

if you press the + sign on the bottom left it should show couple options one of them being "deep research"

I am referring the UI in the web version but it could defer depending on which version you are using.

[craggy granite]

I am on chrome

[sweet merlin]

Here's my two cents since I'm paying for two $20 a month accounts in a overly saturated market. 1st. Temporary conversations, I'm totally against them and anyone in their right mind would be too! 2nd. 7 day limit on attachments in a conversation. Again totally against it, and they should be forever, limit attachments to say 4-6, or no attachments can be uploaded. 3rd. Comet is a bad idea no way around it. There are so many AI companies pushing for their own version of chrome when all they need to do is provide an extension and prompts for permissions. In a market that is saturated of companies like Perplexity.AI why push these bad ideas? Now, for a suggestion. A $40 or $60 a month family plan. In our economy having a plan that allows more than one person to logon with their own private accounts all under one umbrella plan, which has ways for parents to monitor and restrict children's access would be something new and welcomed. Many web based companies like media (i.e., YouTube) and meta have ways so parents can restrict. Heck it would make it more affordable and profitable if small groups of say 4-6 users can use the plan together. It would save them money and give you all more subscribers. Now, I suspect the reason{s) you all pushing 1st-3rd issues is either you all aren't making big enough profits, or you all are seriously loosing money per subscription, if that's the case, these restrictions are going to make it worse because the AI model market is getting very saturated and while you all were the first or second with multiple model access and great searching features, now they are everywhere. Something to think about.

[verbal bronze]

Perplexity Model Council should be available to Pro users as well, just limited LLM selections. LLM council exists already on Github. Making this feature behind $200 price tag doesn't sound right.

[digital cape]

I want to ask everyone: I've been contacting Perplexity's customer support email for over a month now, but they keep responding with an AI agent.

It is just mechanical replies, and not a single human employee has addressed my request yet. What is going on with you guys? Can someone please just have a person reply to my email?

My request is extremely urgent, but so far, no one seems to care. They just keep using robots to stonewall me and shut me out.

I have a question: is Perplexity not planning to address the individual needs of every user?

[red wing]

The current PRO plan is completely disproportionate to the old one, with few features, treating PRO users as if they were free users, totally shameful, and furthermore, the user was not notified and was caught by surprise, a totally cowardly practice from a company that has always been a benchmark.

I expected more from you, especially after more than a year as a user. I never expected this kind of attitude, especially harming paid users, and judging by the comments on social media, it seems that a large portion of PRO users are dissatisfied.

[pine iris]

Hi! We recently started using Perplexity at our company and have been very happy with the Deep Search feature. The recent news also makes us feel Perplexity is still a strong option going forward.

That said, we’ve run into a number of usability issues that make the overall experience more frustrating compared to ChatGPT. We completely understand that, as an enterprise customer among many, it’s normal to receive limited direct support and attention. However, it currently feels like the Perplexity team may not be actively monitoring the #1140622008086970420 and #1358914210339360798 channels.

Is there a better place or process for sharing usability feedback where it’s likely to be seen, considered, and ideally acted on?

[faint trail]

Hi Perplexity, I agree with a lot of the sentiments above. I was disappointed to see an opaque quota for research queries while I pay for Perplexity pro, and finding out yesterday I used up my research quota for the month. I loved the quality of the research responses but I expected a daily quota not a monthly quota that alerted me just as I used it up. I was unaware of before the end of the month and would have done a balance of pro searches and research if I was well aware. I understand you need to make money as a business but I'm unwilling to pay for Perplexity pro if I can't access at least 20 research queries per day, at least because the pro searches don't do a good enough job following my prompts to gather the information I explicitly request for if its not immediately straight forward. Lastly I hope you're not deleting discord messages for feedback that are in good faith because the very next platform where those complaints go are to public forums and the very word of mouth you used to achieve your growth.

My theory, are you about to aggressively advertise Perplexity during the super bowl and in case there's sudden bump in usage you've limited everyone but max subscribers to make costs sustainable?

[inner geyser]

How many people are a fan of online screen assistant ? It's so disappointing that the limits have been throttles terribly in last few weeks

[grave yacht]

Hello, I have a paid monthly plan and I can’t upload screenshots.

[floral raptor]

I allowed notifications once to let me know my question had finished processing, but now Perplexity has started advertising and notifying me about news constantly. It is poor and deceptive by design and not something I think you should expect when you have a paid subscription. I can’t find an option to specifically turn that off; let me know if it is my own stupidity.

[grave yacht]

By the way, with your current paid plan, are you also unable to upload screenshots?

I have the regular Pro plan (not Auto Pro). With this plan, should I be able to upload screenshots, or is that only available for Auto Pro?

[floral raptor]

Like as in images in general?

[cedar iron]

Salut Perplexity,

Il y a quelque semaines, j'ai laissé un message sur ce canal et vous m'avez répondu avec une augmentation de requêtes : 200/jour pour les recherches approfondies et 25 pour les laboratoires (anciennement labs). C'était franchement parfait car je pouvais ainsi faire avancer mes projets et mon apprentissage de développeur amateur. J'ai été gâter et je vous en remercie.
Or depuis quelques jours, comme cela a été dit plus haut par d'autres, vous avez tout ramené à 20 requêtes approfondies/mois. C'est franchement bien 20 requêtes surtout avec le super bowl qui arrive à grand pas mais par mois ??
Faites ça par jour s'il vous plaît, ça permet de travailler, de rechercher correctement, de faire des projets... Surtout quand la réponse à une recherche approfondie avec des détails est : C'est bien ça que tu veux faire ? Un gâchis de requêtes : c'est révoltant !

En plus, j'ai donné des consignes obligatoire à retenir sur tous les threads et je suis en permanence obligé de lui rappeler, c'est éreintant à force de voir qu'il n'arrive pas à les suivres correctement

Je suis d'accord avec beaucoup de messages écrits plus haut tout en comprenant que ce choix de changement est sûrement lié à des stratégies financières. J'ai fais la promo de Perplexity depuis 1an, surtout avec la sortie de Comet et je ne le regrette pas. Malheureusement, ce changement brutal sans prévention du public concerné ne m'incite pas à continuer la recommandation auprès de mes proches

Pouvez vous faire quelque chose svp ?

[half siren]

Hi, yesterday i started to try Perplexity Pro... and even the free chatgpt is not as frustrating as the last two days.... i really dont want to experience the free version if the pro version is not enough aware of its capabilites. I tried to create a label vor some meadbottles and i just wanted a simple beehive pattern. after 20 times explaning how it should look like (and we didnt even get close) the model changed to "Braindead-Mode" and the last 20 prompts where like. "I understand, now i have corrected everything, just copy the code and you can start prining" and then it posts 4 rows of html code. - when asking where the rest of the code is --> "Here it is, just save it" and then its 3 rows of html code .. and so on... just give that zombie the awareness, that there is nothing it can do right now ....

or at least some time frame when Braindead Mode is over ..

[vivid hill]

I guess this has been said multiple times, but current state of pro is a joke.

The new research quotas are completely useless for any production related tasks. If it was 20 per day, it would still be low for some days (but I can see 200 a month working). But 20 per month? That is a joke.

Was previously using it mostly as aid in reversing embedded firmwares and kind of rubber duck, which the search mode is simply incapable of. It provides mostly nonsense and out of context answers.

Yea, it can wind up code for some simple proof of concept tool, but any free tier agent can do that. Hell, I can do that, albeit a bit slower.

If that limit thing does not get resolved before my trial period ends, I will be canceling subscription before any payment is made.

Also, quality of research responses and the amount of aid it provides is definitely not worth the max pricetag.

Sorry for rant, I bet you get enough of that these days. 😄

[ember ginkgo]

I saw a pop-up that "Research, Labs, and Sources" have moved, but it doesn't say exactly where or how things have changed. Please be more helpful when you make changes to the UX.

[ember ginkgo]

Ah, just caught up on the stealth changes to the Pro Plan along with this less-helpful UI tweak. Cancelled my Pro for now. I hate how we've normalized undocumented changes to UIs with web apps over the years, but unannounced claw-backs on what I'm paying for? Not acceptable.

[pine turtle]

Huh?? What’s the point of paying for pro now😭 they just suddenly added massive limits to features I’ve been using since launch. Might aswell not pay at all

Guess I’m switching to Claude

[indigo wing]

Does the discord bot for Perplexity still work?

I added it to my server and it doesn’t work.

[ionic phoenix]

Bonjour
Je ne comprends pas, je suis abonné à Perplexity Pro et mes recherches approfondies sont fortement limitées. Je ne peux plus utiliser Perplexity pro de façon satisfaisante. Quel est le nombre de recherche maximal par jour pour les recherches approfondies pour labs ?

[deep dirge]

Salut, as-tu déjà contacté l'équipe d'assistance ?

@ionic phoenix

[ionic phoenix]

Non

[finite kernel]

Hey Perplexity, Ich nutze seit über einem Jahr Perplexity Pro. Heute habe ich zum ersten Mal die Nachricht bekommen, dass meine monatlichen Research-Anfragen aufgebraucht sind. Das kann doch nicht sein! Ich zahle extra Geld dafür, dass ich unbegrenzten Zugang zum Research-Modus habe! Und wir haben Anfang des Monats! Das ist zudem ein Verstoß gegen den Vertrag! Sie können das nicht ohne weiteres ändern!

[rare saddle]

PEOPLE gotta start contributing votes to the ⁠🌐│product-ideas so they get recognized and help us all

[rough bloom]

Well, I am looking to move my Pro subscription to competitors like Yupp or Abacus AI instead...
These latest changes to the so called agreement is a violation of agreed terms of contract. Why would any business organisation to business on these terms?

For 20 bucks you get X amount of this and this. Next month you realize they cheated you and reduced all rates to 10% of what was agreed upon.

Horrible company to do business with.

[runic plume]

I'm afraid to upgrade because of the goal posts changing as far as usage limits. Give me access to a local model of my own so we don't burn up your servers.

[ember ginkgo]

Expect notification of a class action lawsuit in a few years.

[echo pewter]

So perplexity tells me that I should file a complaint at the BNetzA because the new limits for deep research violate EU law …

What Perplexity should do to comply
To align with EU transparency standards, Perplexity could:[key-g +1]

1. Publish approximate ranges for Pro/Max even if exact limits are dynamic (e.g., “Pro: 15-25 queries/month, Max: 100-150 queries/month”)
2. Show a usage counter in the UI so you know how close you are to your limit
3. Notify users before they hit limits, not just when they’re blocked
4. Document the methodology for how limits are calculated if they’re usage-dependent

[echo pewter]

I give you time until next Friday or I will use perplexity to file the complaint and sue you.

[nocturne plover]

Stop giving pro members such insane limits! I've noticed that first the advanced AI model limits increased, and now the research limits just absolutely got nuked going from 600 to like 20, what the hell?

[echo pewter]

Welcome to the EU. Just have to look up why Amazon had to pay back money to ALL THEIR PRIME CUSTOMERS in the EU 😉

[tribal raft]

absolutly right, I've started looking at altenrtives, this is redicouluse

[elder gate]

JP UI feedback
Translating “Deep Research” as 「深い研究」 is kinda lame. Better to keep it “Deep Research” like ChatGPT and Gemini. If it has to be Japanese, 「ディープリサーチ」 is much better.
“GPT-5.2 思案中” sounds weird too. Just keep it “GPT-5.2 Thinking”.
Also, like everyone else is saying, the sudden limit change is pretty shitty.

[random radish]

How many of y’all use deep research almost always like me? Perplexity…. why in the world would deep research limits go so insanely low out of the blue with no communication for pro members?

[nocturne plover]

I use deep research so often

[random radish]

Yes! I stopped using best and pick your model two months ago. Been running up to 30 deep searches some days up until February 4. Totally disappointed that no Perplexity representatives are chiming in about what’s happening.

[languid leaf]

yo

im really upset

how did i run out of file uploads

when it says

unlimited

file uploads

[random radish]

Is there some massive glitch going on where all pro members are getting non-paying member limits?

[thorn comet]

I'm a Max subscriber. The sudden interruption of performance and patronizing notes from support bots and/or staff this week caused me to finally get on Discord. I nearly cancelled, but I'm quite invested here. This is the sketchiest experience. I've developed expert systems since the '90's for engineers/geoscientists and my investment banking staff over the years. When this system is running properly, it does amazing. I have yet to understand why at times it seems like everything starts performing at 10%. If there was some red light that came on indicating you may be wasting your time - or better yet, the system just locks up so you don't put yourself at risk of trusting results that suddenly turn upside down. Any comments or suggestions would be appreciated. I don't really care to spend a lot of time documenting my issues - I'd prefer to just invest in developing my own system - but it did seem there's a video capture available in Comet now that was pretty easy to use - however, I had zero feedback or communiction to resolve these concerns. The fact I stopped even getting patronizing or acknowledgements of recieveing my troubles by email was very disapointing, especially. as a Max subscriber.

[thorn comet]

I also was prompted to re-sub - and I'm a Max subscriber

[open shale]

The image speaks for itself

[warped perch]

Lmao

[blazing bane]

Damn and you’re having problems too?! Wtfff

[pale gorge]

the amount of RAM this perplexity takes to run Youtube is more than opening 5 youtube tabs in chrome

[blazing bane]

I thought it was just pro- id be PISSSSED if I was paying 200/month and it wasn’t giving 200/month quality

[limpid halo]

@covert sun @native tartan

Subject: Feedback on Reduced Pro Plan Limits (Italian User)

Hi team,

I'm an Italian Pro subscriber and I want to share constructive feedback about the recent changes to query limits.

What I've noticed:

• Pro Search: Limits changed from ~300/day (daily reset) to much lower weekly allocations
• Labs (Create): Now capped at ~25 queries/month, significantly reduced from before
• No communication: These changes weren't announced or explained to existing subscribers
• No visibility: The app doesn't show usage counters or remaining quota

Why this matters to me:

I actively use Labs for [your use case - e.g., "generating prompts and documentation for my AI workflow"]. With only 25 queries/month, I run out within the first week and can't use a feature I specifically paid for.

What would help:

1. Transparent limits: Clear documentation of exact numbers for each Pro feature
2. Usage dashboard: Visible counters showing remaining queries (like ChatGPT does)
3. Advance notice: Communicate plan changes at least 30 days before implementation
4. Reconsider limits: Current Labs allocation feels too restrictive for active Pro users

My position:

I love Perplexity and want to stay subscribed, but these unannounced changes feel like a downgrade of paid service. Many users on Reddit and here share this concern.

I understand infrastructure costs are rising with new models, but transparency would build trust instead of frustration.

Thanks for listening 🙏

---

Apologies if my English isn't perfect - Italian native speaker here

[blazing bane]

Has anyone else had their student discount removed?

[fathom plover]

yea same

[livid prism]

the metadata it keeps updating and carry forwards to all chats is extremely annoying. why there isn't a toggle for this and user control to see & edit what metadata is being passed forcefully to each chat?

[desert field]

Same...

[astral heath]

yeah

it seems

[oblique jasper]

Cancelled my subscription - basically cant even upload a file - Gone from being a pivotal part of my workflow to as much use as a chocolate teapot

[blazing bane]

It doesn’t let me cancel 😀

[fallow ember]

I am currently subscribed to the Pro Plan, yet I consistently encounter errors whenever I attempt to upload any image or document. This is extremely disappointing.
For comparison, I also use ChatGPT Plus, where file uploads work without any issues and I am never prompted to upgrade further just to perform basic actions. I purchased an annual subscription in April 2025, and I do not even use this platform daily I only started using it more actively since yesterday.
If this is the level of reliability provided to paid users, it seriously raises concerns about the quality of the service for free users. Frankly, this has been a very poor and frustrating experience.
Either resolve this issue immediately or provide a clear and transparent explanation of how many files are allowed to be uploaded per chat under the Pro Plan. Paying customers should not be left guessing or dealing with repeated errors for such basic functionality.
I expect a prompt and concrete response.

[zinc echo]

mine seems fine

[ashen fable]

Are all pro users now restricted to really limited research mode limits?
It’s a real pain. I’m finding it hard to find the details of the limitations. I was very happy with the old system. Found it really helpful. I was a big fan of perplexity for it. I told loads of people about the platform. Now, I’m not sure it’s worth it at all.

[tribal raft]

yes i'm looking for an alternate subscription, without the research perepledxity is useless to me, and becuase of the nature of the change with no prior warning, I will not upgrade.
I've tried contacting perplexity support, all answeres i get are ai generated, even when i asked it twice to escalte to a human being.
this is just disrespectfull
if anyone has any suggestions I'm open to hear

[languid leaf]

what do you use research for ?

[blazing bane]

Try Gemini, I’ve been using and love it so far

[lofty kite]

Hello!

I want to say thank you to perplexity team for making discount for students!
I am from Ukraine, and unfortunately cannot afford regular subscription, really appreciate that you guys made it cheaper for me, as a student.

Cheers!

[misty haven]

@native tartan @crimson nymph I've been a Perplexity user for a long time, specifically a "Pro" user, and I've even recommended Perplexity to people I know because of how amazing it is. However, as a "Pro" user (and also as someone aware of the resources and expenses your work entails), I want to express my complete disappointment and outrage at the new updates. These updates increasingly restrict the platform for users who pay for a "Pro" account. I find it frustrating that you've removed "unlimited access" to "Research" and unlimited file uploads (or apparently unlimited file uploads). Not only are some features now reserved for "Max" users, but "Pro" users are getting fewer and fewer features, and to top it all off, we're paying the same price for fewer resources. Furthermore, I found the "Max" subscription price outrageous. You're treating "Pro" users worse and worse; it's deplorable. Furthermore, the prices for higher subscription tiers are extremely exorbitant for monthly payments, and I can only justify them if I think you're targeting Perplexity at wealthy individuals and/or businesses, and less for the average user. Frankly, disappointing. The least we Pro users deserve is to continue with unlimited research and unlimited file uploads (not to mention that, with the new exclusive Max upgrades, we Pro users deserve to have other features (like "in-depth research") also made unlimited, or at least with significantly higher limits, as in the previous version, considering there are upgrades we can't even try). In any case, you're treating Pro subscribers terribly; it's very disappointing, and that's without even mentioning the lack of communication regarding changes and limitations...

[thorn comet]

I spent over six hours today trying to get the MAX out of my Max subscription and came up empty. My expectations of building apps were cut short because systems just don't alllow integration, no matter how fancy the programming seems to be. Now I don't know if I asked the right questions, but Perplexity fell way short in streamlining tasks to Apple Notes and between tabs in Comet today. The other day I had great success having Comet re-wire my Wordpress site. Then that feature was totally disabled for a couple of days after that good experience. I have no idea what to expect from Perplexity Max. I've invested quite a bit setting up spaces so I'm going to hang in a bit longer. The response from support is interesting. They like to let you know they care and check back with you, but I have not experienced any real solutions I can directly attribute to them - other than the system seems to be so sketchy sometimes it takes more to describe problems than they are worth. So it seems at the consumer level, if this is the AI threat everyone is so concerned about, we're safe for a while longer.

[past rain]

Hello, my name is Marios, and i am using perplexity theought revolute premium to have the perplexity pro version, for some time i had it but recently it seems i do not have the perks of pro and i get the option to get pro, does anyone knows something about it? Cause i cannot find a solution to this and i am confused. Thanks

[tribal raft]

Either facts (like politics) or mainly software architecture and design

[misty haven]

Done!

[misty haven]

Perhaps you could contact them at [e-mail]. I don't know if that contact method actually works, but it might be an option.

[heavy perch]

How can you not give an expectation around amount of deep research available? Even a range or approximation is helpful, giving none is a ridiculous answer. How is it that something like that cant be figured out? Surprise, you cant research anymore though you signed up for the service and built workflows to your business around its availability? Are you guys purposely trying to go out of business? You have such a great product you offer. I re-upgraded to Max because I deeply appreciate the product and am fortunate to be able to at this time...but not being upfront with folks who rely on the technology for workflows is really bad for business. Logically, giving a heads up for changes to a plan so people can prepare is also an important step. Would really like you all to stick around and not go out of business from class action lawsuits. Such a bummer.

[wet kestrel]

You’ll get an AI to respond. And the AI will tell you the right things, and so far in my TWO open tickets with Perplexity that’s as far as it has gotten.

[heavy perch]

Yup with no actually useful information or answers.

[wet kestrel]

Hi Perplexity team,
I’m an Education Pro subscriber and my account seems misconfigured in two serious ways:

1. Pro features: I have an active Pro/Education Pro subscription, but I’m being blocked from uploading images after ~60 uploads today and shown “Get Pro” upsell banners, plus “Restore purchases” returns “something went wrong” on both Mac and iPhone.
2. Billing: I’m being charged for the regular Pro plan instead of the Education Pro discounted price, even though my student status is verified via SheerID.
   This is actively blocking my coursework right now.
   I’ve already:
   • Confirmed I’m logged into the correct account on all devices.
   • Verified that my subscription shows as active under Manage Subscription.
   • Logged out and back in, force‑quit/reinstalled apps, and tried different networks.
   Previously, I opened a support ticket that took about a month to get a response and never resulted in a proper follow‑up. Because of that experience, I’m specifically asking for a human agent to review my account’s entitlements and billing and correct them as soon as possible.
   Please:
   • Fix my Pro/Education Pro entitlements so I’m no longer treated as a free user or blocked from image uploads.
   • Correct my billing to the proper Education Pro rate and advise on refunds/credits for the incorrect charges.
   Thank you.

@olive grove

[misty haven]

I completely agree. Besides harming businesses and workflows, it also hurts new subscribers who join the Pro version based on features that aren't real and, worse, are hardly worth paying for—it's almost the same as not having a subscription at all. I really think they should be more honest. If they want to focus on other types of users, they should do it the right way, and this isn't it.

[noble monolith]

#💬│general message

[astral heath]

me too

[hallow igloo]

And me

[random radish]

Signed the petition and paid a few dollars for reach

[willow plank]

When an image is generated, it does not include the "prepared by X" prompt at the bottom. When I run out of Google Nano Banana Pro image generations, it switches to a cheap model without telling me.

[rapid osprey]

Free individual account. When I’m logged out, Finance /finance/AMZN shows the ‘Latest price movement’ timeline under the chart. The moment I log in on the same browser, that module disappears. This makes the logged-in experience strictly worse than anonymous.

[cerulean olive]

Is it a new limitation? If so, should let user know. We can just stop subscription. Just loaded several excel files then this. It was not like that before.

[neon forge]

i am being limited in the number of images i can upload now all of a sudden whats up with that

[cerulean olive]

Everyone got limited. They changed. I feel something fishy going on.

[toxic silo]

hey perplexity team, you are very powerful but what about the limitations your users are getting, cant upload images, i thought the limit resets everyday, but i was fooled, they did only one time limit, means one time limit reached, you will not be able to upload images, cant even self browsing, i literally did only 5-6 self browsing, and now its gone,

I think new ChatGPT Atlas AI web browser is going to take over

[random radish]

Hey Perplexity, what the heck is going on?

Can an actual employee, or better yet, a leader over there, acknowledge all the questions about the Pro plan changes?

Why is this Sam's response after 5 days of waiting to hear back about Deep Research limits dropping as a Pro user?

Here's the reality

[covert trellis]

SheerID approved my faculty status for FREE 12-month Education Pro.
Got "Verification passed, check email", but code lost on Apple relay ([e-mail]).
Verify button disappeared. Support bot replied 5 days ago, no human response.
Account: [e-mail]
Help!

[gloomy kernel]

Does Spaces allow for folks to have a conversation with the breath of all prior conversations in the space?

[languid leaf]

Image uploads being limited is literally so retarded

How

I get deepresearches

But

Image uploads

Crazy

Also Gemini

Pro

Is like $20 and has like 20 deepresearches a day

If I’m not mistaken

[dusty sonnet]

Please help, my annual Pro subscription has expired for some reason, even though it's supposed to end next year.

[weary rampart]

Same problems with annual pro subscription, it’s just disappeared for some reason

[zinc sorrel]

Can we please get space limits increased? 1500 is insanity. ChatGPT is 8k

[little yarrow]

The real insult is the radio silence--For ages, I was the biggest fan of Perplexity. Perplexity is an incredible tool. And yet, the disrespect I've been shown over the last weeks, along with many other customers, is truly unfathomable. Let me start by saying that I think consumers should be understanding— even forgiving - while the businesses they consume from deal with growing pains. Given the quality of the product, I'm not surprised Perplexity is overwhelmed. Yet, what is entirely unacceptable is to slap restrictions on users' functionality without notice, giving incorrect information about why that restriction has been placed, using the restriction to attempt to up-sell customers arbitrarily and without explanation, and then failing to communicate when they reach out for support. Like many Perplexity users, I purchased my Pro Plan mainly for the Deep Search function, which I could use dozens of times a day without problem. Then, out of nowhere, I get told I've reached the monthly limit. Initially, I ignored it, as the restriction message only appeared after several uses, and by the next day, the feature was usable again. Last week, the platform seemed to collapse. I was told at the start of almost every day that I'd reached my monthly limit, despite my subscription renewal happening just a few days earlier. Conversations with the AI support bot were frustrating—being told it was a weekly restriction due to demand, despite the feature's availability varying day-to-day, not week-to-week. The support bot said I'd be contacted by support—this was five days ago. They've also disconnected the phone number listed on Facebook and are ignoring their WhatsApp channel. I'd be willing to give some grace to an excellent product dealing with growing issues and asking for patience. But they should not treat loyal customers with contempt. They should communicate so we have some expectation of what to expect. This is disgraceful treatment of paying customers.

[bronze fable]

sign the petition

[misty haven]

Absolutely agree with you

[little yarrow]

Has Perplexity fully collapsed for anyone else? Just refusing to do anything now - even tasks that don't require browser control - because you've supposedly run out of browser control tasks?

Also, is Perplexity Support also just ignoring everyone - or just me?

[warm merlin]

Did they limit how many searches I can do in research mode? In the pro subscription?

[misty haven]

At the moment it works without problem (I am a Pro user)

[misty haven]

Yes, and much more

[warm merlin]

Bro that’s so wack

Like that’s stupid and looks really money hungry

[little yarrow]

It is not working for me - in fact, the entire system is down for me.

[warm merlin]

Like this is really contradicting

[misty haven]

In my case, I have a limitation in all types of search (except the Pro format), the limitation of file uploads has also been limited, among other problems that other users are having...

[misty haven]

Totally, moreover, the prices to have Max subscriptions are absurd for average people to pay each month

[misty haven]

I'll do a new search to check if it's still working or if it's a general thing

[misty haven]

In my case, it continues to work

[little yarrow]

I guess Perplexity likes you more than me. It could also be that it's restricting some users and not others to deal with whatever internal crisis its covering up.

[misty haven]

Seeing what happened in the last week I already expect everything 😂

[little yarrow]

It'd be funny if it weren't so tragic.

It's the best research AI tool I used - now a collapsed system in crisis too contemptuous of its clients to even tell us what's going on or give basic tech support

[warm merlin]

Yes bro I’ve used it for so many things

But if they don’t revert this I’m ending my subscription this is just absurd the fact that they expect someone to pay 200 a month for this is wild

[dense venture]

(On a Pro plan) I have been seeing this more frequently in my queries which requires pulling financial data. It significantly reduces the value proposition of the plan.

[misty haven]

Exactly...

[swift pivot]

kk

[toxic token]

I'm a Perplexity Pro subscriber and I'm having trouble with the file upload limit.
Yesterday I uploaded quite a few images and received a message that I had reached the daily upload limit.
Today, after uploading only 5 images, I received a message that I had reached the daily upload limit again.

I would like to understand:
Does this limit reset every 24 hours or at a specific time of day?
It hit the limit yesterday around 10 PM here in Brazil.

Could someone help me?

[halcyon heron]

Hey guys!

I have a pro subscription in Perplexity and with a new update it got as a nightmare. No research action in the web version. The Android app limits got messed up, for Pro it gives me only 1 research a day (WTF?). I couldn't find usage limits at all. Unsubscribe from this nightmare also is a challenge

[visual nebula]

This limit doesn’t make sense, I didn’t do much researches and now I’m limited without prior warning, it isn’t a cheap subscription and it won’t worth it to pay for like 10 searches/month

[warped hornet]

As a Pro user I'm FURIOUS about the limits to deep research. That's why I got Pro in the first place. No heads up about the change at all. Is this a joke??

I was such an advocate for your service. But at this point I should just cancel my subscription and use NotebookLM, You.com, Gemini, ChatGPT, Google AI, etc

[peak tundra]

Hey, guys!

I find the current perplexity deep research so under quality, the responses are so structured, feel like overly AI-generated, and the overall gist of the answer seems lifeless the results are poorly presented, and too many bullet points, the overall quality is unsatisfactory

And also the quota limit

20 per month is ridiculously low (for Pro tier)

Bring back the old deep research

Or at least keep a feature that allows you to use old deep research, it was really good

[odd pond]

hi there

is your research showing like this in the UI?

[frank cedar]

Totally. I thought it was a joke when i read itfirst. Before we got 600 now just 20 and the redult are not 300 times better. Not even twice as good. If this dosent change back soon, im gone.

[frank cedar]

Same

[dire bridge]

I’m flabbergasted at how my 240$ a year pro account is massively limiting my Deep research searches. Even more infuriating is that this change was made without notifying paying customers. I really hope the team gets it together or I will leave the paid account and go to another provider.

Has anyone gotten clear answers about the new limits? Or does anyone have a contact at the company.

What a disappointing customer experience

What BS

[little yarrow]

PRO USER GENERAL FEEDBACK PLEASE:

Can you confirm that, before the last couple of weeks, limits on features like Deep Search and Browser Control were far higher than they are now?

Personally, I used Deep Search for work between 6 and 12 times a day before these unannounced limits came into effect. Was this your experience too?

[misty haven]

Yes to both questions

[slow sage]

I use it for my tutoring company. I make curriculum and comprehensive guides. I would use deep search over 30 times a day

[flat latch]

ñ

[rich plinth]

Yes, Deep Search capabilities have been decimated, right in the middle of my thesis project, this is creating great harm on my academic performance since the scope I chose was based on this technology

[finite kiln]

Google Gemini Deep Research offers more limits than Perplexity - using Claude Opus 4.6 is the wrong choice. The best feature Perplexity offers is Deep Research from my perspective, and it needs to offer minimum 10 Deep Research queries in a day and not in a month. BTW, Perplexity is not offering a public figure for this either - how much is the actual limit?

[odd pond]

20 a month

[finite kiln]

20 a month for pro subscribers is nothing.

[odd pond]

that's logically wrong statement tbh

[cedar iron]

C'est quoi ça Perplexity ? @covert sun @native tartan @south ether
Ce n'est pas la 1ere fois que ce style de message arrive. Comment voulez-vous que l'on reste avec vous pour nos projets par exemple si on a ce genre de proposition chaque semaine avant d'être bloqué ?

Si nous prenons Pro uniquement et pas Max: c'est à cause du prix de Max !!
Personnellement, je ne peux pas prendre Max à cause de ça et toute c'est restricton me bloque dans mes projets !! Si malheureusement il y a encore ce genre de restriction, je vais malheureusement aller voir la concurrence comme Mammouth.AI, européen qui ont moins d'option comme Comet par exemple. Et pourtant j'adore Comet !!

Pouvez vous faire quelques choses s'il vous plait pour assouplir ces restrictions ? Je suis en plein travail et déjà plus que 2. Sauf que j''en ai besoin

[strange yarrow]

Hi everyone!
I’m a user from Russia and I’m planning to buy Perplexity Pro via the App Store on iPhone.
Support told me that due to OFAC sanctions access from Russia is not guaranteed and may be blocked even after purchase.
Can someone from the Perplexity team please confirm if there is any safe way for Russian users to use a paid subscription without the risk of losing access?

[cedar iron]

And now it's down

[cyan sedge]

Subject: Voice Input Mode — Move the Cancel button away from the Submit button

Device: iPhone 17 Pro (small form factor)

Description:
In voice input mode (speech-to-text for composing a prompt), the Cancel button (square icon) is placed immediately next to the Submit button (checkmark icon) on the bottom-right of the input bar. These two buttons have opposite effects — one sends the message, the other deletes it entirely — yet they are only millimeters apart.

Problem:
As a power user who relies heavily on voice dictation to compose long, detailed prompts (often 3–4 minutes of continuous speech), I have accidentally tapped the Cancel button instead of Submit on multiple occasions. This instantly erases the entire dictation with no way to recover it. On a smaller iPhone screen, the proximity of these buttons makes this a recurring and extremely frustrating issue.

Suggested fix:
Move the Cancel (square) button to the left side of the input bar — for example, next to the language selector (“Fr”) — or add sufficient spacing between Cancel and Submit.

Alternatively, add a confirmation dialog before discarding a long voice input (e.g., “Are you sure you want to discard this message?”).

This small layout change would dramatically improve the voice input experience for users who dictate long prompts on smaller devices.

@crimson nymph Hey, is this the right place for UI/UX feedback?

[crimson nymph]

Yes this is the right channel for that 👍

[shut arch]

@coldicemonkey muted

Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds

[wind kettle]

Honestly given how Perplexity has been functioning these days I would say don’t bother. There are better platforms out there.

[frank cedar]

You can clearly see that the limit of 20 deep researches comes from this partnership with claude.

While using Claude Pro i runned daily into the set limit from claude and it made me furious to pay for such bad service.

At this time i was always an advocate for perplexity for this amazing limit of 600 deep researches.

Time to look around for a better service. Where do you guys head to? I like Gemini Deep Research and Manus abilities. But both take too long to create answers.

[faint spire]

What’s up with kimi dr

(Don’t know shit about it just curious )

[bright sonnet]

The pro plan is effectively worthless now that deep research has been functionally removed from the plan. The max plan is terrible value and right now people are going to switch to Claude cowork/code or codex plans and abandon perplexity before they upgrade to perplexity max. Perplexity deep research will probably never be able to compete with the longer format deep research products like Google or Manus. The previous version of deep research is the only part of your product worth paying for and now there’s not any value with so many other basic llm search options and longer format deep research products. It’s time for monthly pro users to cancel their subscriptions.

[cyan sedge]

HOLY fucking shit you guys already updated the app

[sullen surge]

It got so much worse. Since the new update, they moved from a transcription feature, where you could talk for several minutes, take pauses, brainstorm and think about your question while saying it, to a live transaction which automatically ends the listening when you make a pause. Even if you just breath for a second, perplexe stops listening to you, sends the prompt, just to tell me that it’s unable to help me because of missing context. IF YOU WOULD‘VE LET ME FINISH… I would have provided that context. I loved perplexity for that feature, which Gemini already switched away from. But if @shut arch won’t give us at least the option in settings, I will switch to Claude, the only place where normal speaking is still possible

Luckily I have an iPad where the perplexity app is still running the version before the update. I rushed to settings to turn off auto-updating, but I have no idea how long the app will be able to run without any updates

[sullen surge]

Just to help you with your problem, that square button actually is not a cancel button. It just looks like that. If you tab it, you will end the dictation, but if you wait a few seconds (roughly 5 seconds per minute of talking) your voice is still being transcribed and actually placed into the text field, you can edit and send it. I actually use that button all the time, cause while the transcription algorithm is really good, with names or technical products it still often misunderstands, what I’m saying

[bronze thunder]

I can confirm this. In December 2025, the deep search limit is 600/day, in January 2026 it's 200/day, now it's 20/months

[ruby scarab]

What's happening with the context window. It loses track of things only a few questions deep in a thread. This is for Pro and it was definitely not like this even several days ago.

[faint spire]

I’ve had the same issue, just… forgets what planet it’s on and what thread

[deep palm]

I've been using Perplexity for four months, and it was excellent. However, things have changed since I started, as the limits have been reduced and uploaded files have been restricted. I'm concerned that good providers may be changing their policies to generate more revenue.

[jolly kestrel]

Yeah happened to me as well, it's like I'm talking to my grandma now, can't remember what's been said 2 minutes ago...
Since all these changes to the limits, it's all going downhill. I'm cancelling my subscription, will look at the competitors for a good replacement! Probably going to end up with Claude!

[agile notch]

Can I have my money back. At least the last month. Because I think that with this limits, I'm paying pro for nothing (they promised something else)

[odd frost]

Hi community,
I mainly use Perplexity for scientific analysis and find the Deep Search feature extremely effective.
I use it reasonably and in the past I have never experienced any restrictions, but recently I have been receiving messages saying that I have reached my usage limit.
Has Perplexity changed its quotas? I find this very frustrating, as I am paying the same price (I have a pro subscription) and the service has been reduced.

[neat sleet]

Yeah, perplexity changed its quotas recently, many are mad about it, its 20 a month for pro users for depp search now

[jolly kestrel]

BUT you can do as they suggest everyone who ask about it: "upgrade to MAX for the same amount (almost, it's only 10 times more) !"

[odd frost]

@jolly kestrel & @neat sleet Thank you for your response.
That's quite a limitation! I easily used 2 to 3 deep searches per day... I understand that in the middle of my work, I can't get anything done 🙂 . It's really annoying.
And clearly, I'm not willing to pay $200 for 50% more searches (which would still be insufficient).
I hope Perplexity will reconsider the situation...

[mighty basalt]

When we getting GLM5

[neat sleet]

Glad I could inform you a bit so you know what's the situation rn.

[cyan sedge]

You’re absolutely right, and I hadn’t realized that. That’s really good to know. I was treating that square button as a cancel button this whole time.

And honestly, I’m really happy with the new change overall. At the very least, I no longer have the problem where I’ve been speaking for a full minute straight without any pause, only for the entire message to suddenly get wiped out. So that’s a big improvement.

That said, my main point isn’t really about the transcription service or live translation itself. I just wish the transcription would end when I decide it ends. If I take a few seconds, or even 30 seconds of silence while I’m thinking, I don’t want it to stop on its own. All they’d need to do is subtly adjust the behavior so that it doesn’t automatically cut off. It should stop when you’re done speaking, and being done speaking means you’ve hit the “Send” button. Not when you’ve been quiet long enough for their liking.

But thanks for the clarification, that actually helps a lot.

[tribal raft]

did anyone get an offical response from perplexity? not from the anoying ai?

[deep palm]

Perplexity Pro feels pretty useless now. I think we should all cancel or they need to go back to the old model. It's frustrating when companies seem more focused on making money than happy customers. Without any announcements.

I'll leave a bad feed back every day until my subscription ends

[deep palm]

Yes and now they ruined everything

[boreal stump]

Bro they r Making it shit day by day... It's Useless yae , Chatgpt gives u the same thing per day... limits on uploads & shit , it's fucking useless , feels like free version

can't even upload files or images , they limited that thing too... this sucks , free version of Chatgpt is much better then perplexity pro not gonna lie

[deep palm]

Yes and i got a timeout for leaving my true feedback im cancelling and leaving here anyway I'll just use my own self hosted models

[boreal stump]

u got 1? damm

[boreal stump]

& even i tried to email them bro , that fucker sam AI bot ... fucking replies everytime even after asking for human staff

can u tell me how to even talk to human staff?

[deep palm]

Idk tbh is just useless yall should not buy anymore and they should think of their customers

I already cancelled my subscription

[boreal stump]

bro even chatgpt refresh's ur upload of files every 24hrs & for queries

yae , I'm gonna cancel it too

[rotund epoch]

Enshittification

[digital warren]

Cause mostly they are hungry for money

[boreal stump]

yup , i said it in general chat lol

[digital warren]

Daam i didn't see it there it's just my thoughts hehe

I mean it tho cause when I first seen the pro subscription which they offering it's feel so better unlimited research and i think labs used to be 100 / month those were the cool days it's been almost 2 years using perplexity now it feels so useless so much I think I future they will limit the normal models chats too

[boreal stump]

They already limited the no of files u upload per month , from few days ... Idk exactly how much u can upload... & Yes I think they will do it .. but it's already worse . No one gonna buy pro with these shit features offered , the pro features perplexity is giving are already free features in chatgpt

[digital warren]

And I aren't paying 200$ trusting these people when they can't even keep there word for the pro version

[boreal stump]

But it refresh's after every 24hrs or week

[boreal stump]

It's a waste man 😂 don't pay yae.

[digital warren]

I mean I did see the words sayain they can do changes in the plan whenever they want but this much sudden changes feels too much off and they did so many changes but they doing same promotions to this day stating they giving unlimited shit 🥀

[boreal stump]

Well changes when they want yes, ... But removing most of the things what feels like free version of chatgpt .. y pay fucking extra money for these shit things when u get them for free in chatgpt free version & better things if u pay for chatgpt itself directly

[digital warren]

I think the thing happened here is they promised alot of shit and they given too much now it's getting hard on there backs and the gpu and ram hikes they thought they could sneak some shit out from the paid users too

I mean perplexity users are mostly power users that's why people noticed the shit pretty quick

[boreal stump]

Nah it's the 2nd & wost limiting they r doing...

On 1st limit they did decrease the size. Of search's & few things

Per month but now they limited it too much

Basically free version of chatgpt

[digital warren]

Lol it's even bad bud

[boreal stump]

Yae true... But there is no other llm worse to compare 😂

Of limiting it's features

[digital warren]

I mean idk I feel like claude doing great these days and gork too

[boreal stump]

Yae... U can buy pro version of GPT. Then buy pro version of perplexity

[digital warren]

I have both 🥲

[digital warren]

I mean I have gemini too ngl

[boreal stump]

Lmao

[digital warren]

What can we do nothing is perfect in life

[boreal stump]

True.... & Nothing is permanent in life too

[deep palm]

Yes no one will buy max too because they did that to pro they think we will buy max ain't buying again

[cedar iron]

par contre c'est ptn d'abusé là !! Toujours le même problème à vouloir forcé l'upgrade vers Max. Pensez vous à ceux qui ne peuvent pas et qui ont besoin de travailler avec les différents modèles ? @covert sun @native tartan @south ether @shut arch Pouvez vous faire quelques choses s'il vous plait ?

[sweet merlin]

Perplexity.ai better reverse course or there will be a exodus and than they'll be sunk! There are more and more multi-AI model plans with research out there, so their going to find out quick that the consumer is the boss!

[sweet merlin]

Yeah they changed it by a lot. Now it's down to 20 / month unless you want to spend $167 / month but who knows when that will restricted too. Also uploads have been restricted to 7 days before autodeletions, and even the number of pages per uploaded pdf is down to 9 pages per .pdf. They are burning their customers without warning, or asking for feedback! I guess they want to see how long it will take for them to loose all of their faithful customers because these suddenly restrictions are making enemies not friends.

Time for all of us to go to review sites like TrustPilot Reviews and let our voices be heard! This sneaking crooked crap is going to cost them! Update! They are already a 1.5 out of 5 and falling.

[cedar iron]

pareil : j'ai besoin d'analyser des images et documents svp !

[sweet merlin]

Just to let the Perplexity.AI team know a little about what their underhanded changes have cost. I am a caregiver of a woman who is disabled, who has been using Perplexity.AI to help her work on getting reclassified in NYC Workers Compensation. She has been happily working on this for months since we moved from OpenAI to Perplexity.AI, but now out of the blue she is unable to get the help she was receiving just a couple of weeks ago! We are both on a fixed income and we could barely pay out $20 a month each, and felt it was a god send for the services you all had given us, but now you took them pretty much away and put them behind a $167 a month for each paywall! No Announcement, no email, just flipped a couple of switches and BAM no more than 20 deep researches per month, a more limited number of files that can be uploaded and reviewed per conversation and each file now has a 7 day timer! We are looking for alternatives and have put our negative reviews on TrustPilot.com and we are telling all of the other Perplexity.AI users who were screwed too, to do the same and to go to other review sites and put negative reviews there too! What the team here did maybe legal but it was completely unethical and they should be review bombed for it!

[arctic prawn]

@shut arch @covert sun @native tartan @south ether Hello Perplexity Team I have a request for you. You should increase the number of searches and other features by doubling them per month for pro subscriptions, and make the maximum features available to pro subscriptions (limited, for example, by dividing them by 2 or 3, but still accessible). Keep in mind that OpenAI has almost no limits in its $20 plan.

[sweet merlin]

They had until this month given us 200 deep researches per month now 20, file retention was also 30 days instead of 7. They just screwed everyone in the Pro subscription level with no warning or announcement. Some people who have yearly subscriptions were completely screwed and cannot get refunds. If you are new Azertix, best move on to another company. The team running Perplexity.ai do not believe or understand good customer support.

[odd frost]

My annual subscription for my pro account is about to expire, and I will not be renewing it (unless @shut arch corrects its error quickly).

[strange oasis]

Finance Search Option has been removed recently. Why???

[faint spire]

Cuz if you’re gunna fuck things up you might as well go big

[deep palm]

Worst and most useless AI I've ever dealt with. I got limited after uploading 5 images. DeepSea allows 50 images per upload, which is free. And on top of everything, Perplexity is still quite after hundreds of negative feedbacks, thinking people will buy like before.

[torpid ravine]

👍

[boreal stump]

not sure what to say anymore

[charred orbit]

Incognito should have an “are you sure?!” Before closing .. oops.. i usually SHARE the url of incognito chats.. this was a long one i closed and it is now gone by accident

[boreal stump]

those r google drive share links directly , not incognito onces

[cursive oak]

@shut arch please fucking reverse these changes or you will bancrupt.... this is not normal every single site is better than this

[cursive oak]

try scira ai

if you want multimodel

[jovial trout]

And if I want save a chat into a file, which the file can be referenced by the Space?

[grand hound]

Deep research and labs limits for pro users have been reduced?
Previously, deep research is limited to 300 requests per month and now it has gone to 20. Labs was 50 and went to 25.

[odd pond]

correct:
#👑│pro-lounge message

[lavish swan]

If you can make it so that under the tune option you could set a temperature range of how creative it can and branch out with the sources with the searches for any pro, web or research search

[river plover]

I think perplexity pro isnt what it used to be. Max is the old pro, but for 10 times the price. U didnt keep the value of pro.

Does anyone know any alternative to perplexity? I will plan to change in the future. Right now even if I pay for 2 AI AT ONCE it’s still cheaper than perplexity max and it brings A LOT more value. (Chatgpt and gemini)

[grand hound]

I have a lot of affinity with Comet from Perplexity. I prefer his response style. But unfortunately with these points that you brought, I'm considering Claude + Google

[grand hound]

I just noticed that the student plan (education) has been removed. I can't find it anymore

[vague cargo]

It's ridiculous that Perplexity is expecting us to pay $167 dollars for more than 20 uses of their reasoning models. That's not affordable to consumers in this current economic climate. I've been a big fan of perplexity its sad to say that I will not be using it in the future.

[sharp harness]

Pro android user

I hope for a better audio user experince in noisy environments. Even the slightest noises in a room will prompt perplexity to interrupt itself in its response and start anew. Also a slight echo when speaking will prompt it to immediatly give a response when I'm in mid sentence.

Great app all around, been using it for about a year now.

Thanks!

[frank cedar]

Just wanted to let you know guys how "amazing" the new deep research from perplexity is (that forced Perplexity to reduce the limit to 20 deep researches per month). I used Gemini Deep Research to gain information about a topic and asked Perplexity Deep Research to do fact-check on it.

Answer from Perpexity was, that it couldn't find some information that was stated in the Gemini-Paper.
I went to Gemini to check the source, which was 100% accurate and gave that source to Perplexity and asked again. Perplexity was then able to finish the fact-check and could admit, that everything was correct on the Gemini paper. So i used 2 Deep researched to check 1 Deep research from Gemini.

[signal grail]

I’m so disappointed that the feature I use the most, the “old Labs”, in Perplexity was cut down to just 25 queries a month.

It’s really not worth keeping Perplexity Pro anymore. I guess the party’s over, the hype is fading, and reality is setting in.

The bubble’s about to burst.

[fierce forum]

hi to all great PPLX team!
IMPORTANT FEEDBACK:

please, create the chance to copy threads without citations! I would avoid to be forced to use an invasive third parts plugin on Comet to do it. is very useful can have this chance.
PLEASE

[tame valley]

can we please implement a toggle to stop comet from taking over the browser when using perplexity, as it stops using it's mcp integrations with Gmail for example and goes into browser mode.

is there a way to make it only use mcp integrations

[dark fjord]

as of today: iOS has read aloud but the Mac OS app lost it recently. Any official update?

[wintry widget]

I was an avid Pplx supporter and paying Pro user.
Going to switch away if I find anything that competes.
Looking for accurate answers and low political bias.

Pplx has been super strong on both points, but that's now literally the only edge they have. :\

[sterile leaf]

I just want to use upload unlimited files! I love pplx and comet! but this is seriously making me so upset!

I have pro and this is just so heart breaking

[noble monolith]

please use https://discord.com/channels/1047197230748151888/1047649527299055688 @ornate haven @languid hound

this channel is dedicated for feedback about Perplexity, thank you

[tranquil spear]

UI problem with Announcements
The announcement of task was in the side left plane and is now moved to the right top corner. That is not helpfull, but if it would work then it was workable. When you click on a announcement or start a thread the bell of the annoucement is not visible anymore --> this seems like UI mistake, would it possible to solve this?

[bleak gale]

Unresolved Referral Credit Issue – Seeking Help from the Community

Hey everyone. I've been trying to resolve a referral credit issue with Perplexity support for months now with zero resolution, and I'm hoping someone here can point me in the right direction or that this reaches someone who can actually help.

The situation:

I earned referral credits through Perplexity's own program — my account even displayed a running total of months accumulated (14 months of Pro). At some point, that page was silently removed and I was told the cap is only 6 months, which directly contradicts what was shown to me when I made the referrals.

What makes this worse:

Perplexity is currently running the exact same promotion — "Get up to 24 months of Pro. Earn 1 month of Pro for each friend you refer." I have a screenshot from this week showing it live. So the "different program limits" explanation I was given doesn't hold up.

Where I'm at:

• 13 successful referrals confirmed by Perplexity's own support agent
• Months of back-and-forth emails, repeatedly transferred, no resolution

Last response warned me that following up would push me to the back of the queue (a support tactic that feels designed to discourage persistence)

I'm not looking to cause drama — I just want what was clearly promised and displayed in my own account. If anyone from the Perplexity team is here, or knows the right channel to escalate this, I'd really appreciate it. Ticket reference: processed-7B6CF50B-68D6-4BFF-909D-C8A3D2D0D07D

[proper lily]

Legibility Issues
Hi, the output font has changed to an "editorial style" serif font, which makes it very hard on the eyes and difficult to read. I tried different browsers (Chrome, Safari), and the illegible serif is everywhere for outputs. Editorial fonts make sense for print and more long-term body reading (not this specific one tho, its not legible), but sans serifs are much more legible for on screen purposes and smaller text sizes. Not sure if there was any AB testing and research done prior to the change, but it would be great it there was at least an option to switch back (or between). Its extremely hard to read and it disrupts my daily workflow. Currently looking for plugins or add-ons to change the font, please let me know if there is a solution.

[indigo flame]

me and the bot agree at least let us choose the font if you're gonna make it ugly

[grand hound]

What happened? Do you have any new recent limits?

[sterile leaf]

I use many files daily, i take pictures of my computer from my phone, i took 4 photos yesterday and it said i hit my limit…

I never had any problems, but when I was working on sunday night. I got through 15 photos before it cut me off.

[grand hound]

Pro user?

[sterile leaf]

yes sir I am

[astral birch]

I’m sure this has been said many times over, but the web browser font is extremely difficult to read. It’s almost illegible to me.

[brave marlin]

Hey everyone in #1111767882007842916-support or #billing, anyone else dealing with delayed responses on unauthorized charges? I've been waiting 5+ days on an email to [e-mail] about a random $20 hit from Feb 12—no reply yet, super frustrating.

Need tips on escalating for a quick refund/investigation without sharing details here. Who's gotten fast human help? Thanks!

[blazing estuary]

Pls make perplexity api from vscode 🙏

[compact sapphire]

I noticed that "Discover" is not an option as an Android Shortcut anymore. Anyone knows anything about this?

[teal dock]

@shut arch
Dear Perplexity Team, What is the point of expanding AI models if you then reduce the usage limits? On top of that, most of the models you add for the Pro plan are completely free on their official apps.

Honestly, the current subscription, with its present limitations, offers little value for the price. It simply does not justify the cost.

[grave niche]

@shut arch You guys should support Vietnam, thank you.

[flint laurel]

+1. It’s not just hard to read, it’s also annoyingly ugly. Who was drunk enough to think it was a good idea to change the font to this? At first, I thought I had accidentally installed an old LaTeX distribution and was using the default template. Please change it back or let me choose the font. This is not bearable.

[grand hound]

@native tartan @shut arch please, help us about new font in Comet. Today the font change to Serif. Is it possible to configure to go back to the previous font family?

[sharp cove]

So tired of the caps now in place. Hit them every weekend as i do my projects. Hit deep research cap: much of it iterating with it since it does t get it all right up front. Hit pro search limit yesterday for the week! Hit file upload too; mostly uploading screenshots to trouble shoot tech setups. Never hit these before this month. Has me looking around.

[winter knot]

Hello, I am a Perplexity Pro user since two years and I never joined the Discord channel. I decided to do it today just to say that the new serif font on web UI sucks. It's hard to read, after a long read my eyeballs sore. I can make Perplexity generate a Tampermonkey script to revert back the old font, but since I love Perplexity so much, I'd like that the team would implement an option to choose the preferred font (and possibly don't use this as default) for all people who don't even know what Tampermonkey is.

[frigid gull]

Is there anyway to turn off Discover Notifications on ios? Perplexity Pro user -

[ebon badger]

Has the font problem been solved? It looks terrible

[radiant oyster]

@shut arch Is Perplexity planning to integrate Grok 4.20 into Perplexity?

@native tartan

[blazing wolf]

it's been months and no reply from @toxic forum

[lethal lance]

Just came here to say I'm disliking the new font. Apparently there is no setting to change it?

[runic sand]

On what basis do you change a core aspect of the UX like the default font without informing users in advance, explaining the rationale and expected benefits, and providing an option to keep the previous appearance or choose an alternative?
For a tool that people use for hours a day, typography is not a cosmetic afterthought... it’s part of accessibility, comfort, and productivity.
I just cant read this font, it slows me down, and no option to change is ridiculous

[surreal mica]

This is terrible for @toxic forum , 20 deep research queries a month for Pro subs, you've got to be kidding me. Looks like it is not replacing google for good. At this rate, company bout to go into oblivion, with google and OpenAI already catching up.

[ocean jacinth]

Damn, what's wrong with font? Absolutelly horrible shit, any way to change it?

[stuck glen]

Hey, everyone! I need your help!

Since yesterday, some weird bugs have started happening.

I choose the Claude GPT model or any other, in a new chat or an old one, it doesn't matter.

But after sending an SMS, IT AUTOMATICALLY changes to the “best” option, even though I don't need it. Before, I used the one I chose. I have a $17 subscription, not a free one. Everything was fine before. Am I the only one, or what kind of update is this?

Also, if I write choosing claude or gpt, etc., and ask what model it is, it answers me - grok!

[sleek mountain]

anyone to assist on student subscriptions? got it canceled and now can't resume it back

[noble monolith]

please contact [e-mail]

[whole root]

Perplexity Team. Over the past few months, the overall value of the platform has consistently declined from a user perspective. The way you started switching models on users without notice was bad enough already, but with the recent implementation of an extremely low and opaque cap on Deep Research usage, while simultaneously marketing it as “high volume” access for Pro subscribers, is deceptive and anti‑consumer. Presenting vague “monthly limits for average use” without disclosing concrete numbers, and then enforcing a cap after only a handful of Deep Research runs, is not compatible with transparent or fair dealing with paying customers. (I apparently reached my February cap with just 6 queries). As a paying subscriber, I relied on your public representations regarding the capabilities and usage limits of Pro. The current Deep Research cap feels like a classic bait‑and‑switch: the feature is advertised as a core benefit of the plan, but in practice is so tightly restricted now that it is practically unusable for any serious workflow unless the user upgrades to the next plan. For these reasons, the new Deep Research policy is the final nail in the coffin for my use of Perplexity Pro. I have cancelled my membership and will be advising others to carefully review the effective limits and recent changes before subscribing.

[formal thorn]

I left. Pure and simple. Got the basic Gemini for queries and Claude Code. Too bad. I really liked Perplexity.

[fierce forum]

I agree but only in part. I think is also true that they are and continue to be always the leaders for researches, study, and many other tools for manage up-to-date informations and a good coding agent for little projects.
True, is not so fair that Deep Research its passed from 600 daily to 20 x week without inform Pro users before, but they also EMPOWERED much Deep Research and now is much way powerful than before and with more features than before cause now is using a good agents orchestration.
And this costs much more resources for AI servers, so is technically impossible for Pplx continue to give 600 daily usage of this empowered feature for the same Pro subscription price.
But I agree to ask to PPLX to increase a little the actual weekly cap of 20 Deep Reseach.

[worn shuttle]

In the recent weeks I've noticed a weird issue where PPLX says there's a regional block for pretty much basic images such as a logo of a PC, or even an image of a dog, and I found myself generating up to 4 images a day.

What do I have a pro account for if I can't even generate the basic images with it??

Also, now I seem to not be able to ask for help as everything counts as "research" and you get to the limit VERY fast.

What's up with that?

[finite kernel]

Its monthly. Not weekly. So its not usefull anymore.

[sharp cloud]

Feature Request: Unified Library in Sidebar with Cross-Space Search & Filtering

Hey team -- loving Perplexity overall, but the current chat history experience is pretty fragmented and I wanted to share some specific feedback.

The problem:

• The sidebar only shows a small subset of recent threads, not a full history
• Threads created inside Spaces are siloed and don't appear in the main history view
• To find an older thread, you have to navigate to /library separately, and even then you can't search across Spaces from one place

What I'd love to see:

• Library accessible directly from the sidebar -- not just a link, but an expandable/scrollable view of all threads
• A unified search bar that searches across ALL threads including those inside Spaces, with filters to narrow by Space, date range, or search mode (Pro, Research, etc.)

The building blocks are already there -- the Library page has search and filters, and Spaces are great for organization. It just needs to be unified into one seamless experience.

Thanks for considering it!

[grand hound]

@shut arch please remove this new font with Serif. It's terrible to reading articles and searches.

[teal cove]

@shut arch something you used to have as a feature that has since gone away that I would REALLY like to have back is being able to see what language model was used for a prompt I submitted.

It’s very interesting being able to compare and contrast the outputs of various models to the same prompt, so could you please make that a feature again? Otherwise, I’d rather just pay the individual companies to use their models individually and not pay your company so I can know what was used for what prompt.

Request: please make it so once I’ve submitted a prompt and received an output from a selected language model that I am able to see the name of the model that was used. For example, Gemini 3 Pro or ChatGPT 5.2 Reasoning. Thank you.

[gusty wadi]

@shut arch I have Pro subscription from Revolut, but unfortunately something is wrong from a few days. The app and desktop asks me to upgrade to PRO. When I trying to recover, I get information, that something went wrong

[novel tulip]

Hi guys I am using Perplexity Comet browser, I am not able to give access to my browser to perform the actions for me.
I am a non-pro user. But I saw this feature in Comet ad and YouTube videos.
I searched everywhere in settings but it's not working, it's not taking control of my webpage. It's just writing down the responses.
Can anyone help please? I am frustrated. I am a learner and it is hampering my studies.

Please help @shut arch or @anyone
I will be very thankful to you 👏

[grand ember]

How can I renew my pro status through Revolut …

[paper idol]

I’m writing as a paying customer with serious concerns about the lack of transparency and auditability around Deep Research limits.
Right now, there is:

• No per‑day or per‑query usage history available to me.
• No way to see when or how my Deep Research quota is consumed over time.
• No persistent, account‑level record tying my current subscription to a clearly documented Deep Research allowance and reset schedule.

Because of this, I have no way to independently verify that:

• The limits I’m receiving match what is advertised for my plan.
• Those limits are not being changed or reduced mid‑cycle.
• “You’re out of Deep Research” prompts are correct rather than the result of a bug or silent quota change.

As someone who depends on Deep Research and is willing to pay for it, this lack of visibility directly undermines my trust in the product and makes me reconsider whether to continue my subscription.

I’m specifically asking for:

1. A per‑day (or at least query‑level) usage view or export for Deep Research tied to my account.
2. Clear, stable documentation of my plan’s Deep Research quota (exact number of runs and the reset window) that I can verify against my usage.
3. Confirmation that my current Deep Research limits match what is currently advertised for my plan.

Without better auditability and quota transparency, it’s hard to justify continued paid use. Please escalate this feedback to the product team and let me know what can be done in the near term.

Thank you.

[thin swan]

Wtb vibe switch

What a nerd

Am I hacking the Gibson now?

https://tenor.com/view/anonymous-hack-the-planet-hackers-hacker-hacking-gif-14778475

[deep flicker]

#1111765634267742408 Pro renewal + ToS “personal, non-commercial” — I feel blindsided

Hi Perplexity team 👋 (and everyone),

I’m a paying Pro user and I just renewed my annual subscription. Only after the renewal I noticed the Terms of Service (last updated Jan 23, 2026) include language that restricts usage to “personal, non-commercial use only” (and also bans “commercial advertisement or solicitation”). https://www.perplexity.ai/hub/legal/terms-of-service

Honestly: that feels like a gotcha.

Perplexity Pro is marketed and perceived as a tool that many professionals use for real work: research, drafts, client-facing summaries, internal notes, sometimes even copy concepts. I’m not talking about reselling the tool or abusing it — just normal professional usage in day-to-day work.

So I have a few direct questions:

1. Is Pro really intended to be non-commercial only?
   If yes: why is that not clearly communicated at purchase/renewal? This is a major decision factor.
2. If this is about specific high-risk use cases (ads spam, bulk automation, reselling), could the ToS be narrowed instead of effectively saying “no business use at all”?
3. What are my options if I do not agree with these updated terms?
   I literally just renewed for a full year. I don’t want to be stuck paying for something that I can’t use for my normal professional workflow.

If the intent is: “Business usage belongs in Enterprise,” can you clarify what Enterprise actually unlocks legally, and whether there’s an upgrade path (and how you handle recent renewals)?

I’m writing this as constructive feedback because I genuinely like the product. But the way this landed — quiet update + restrictive wording + right after renewal — really kills the vibe.

Would appreciate an official clarification from Perplexity staff here.

[cerulean olive]

Better don’t expect anyone from the offical channels would reply, just stop renewal, perplexity has been getting rotten with greeds. They can adjust the usage limits but not significantly reduce to almost like nothing. I renewed in Dec 2025. From now on, never do yearly subscription with any SaaS that doesn’t have any track record.

[fierce forum]

with filters to narrow by Space, date range, or search mode (Pro, Research, etc.)
and filter also for the temporary threads please

[inland slate]

Perplexity is becoming worse over time.
Usage limit and Terms of Service changes without asking the user.

You are acting like thiefs, who stop service after takeing the money.

Trustpilot tells the whole story of this bait tactic.
You are lucky that you are not operating from the EU.

[open carbon]

Im here because a big youtuber just posted about your new TOS and how it basically makes the entire thing moot for smaller single person work related use. Between forced attribution of perplexity use and no commercial use at all, not even to help with advertising or yt headlines or anything, its essentially scrapping your whole service compared to other places that allow for that. Im just saying as a customer Im unhappy and I wont be using it unless theres a clarification or adjustment.

[old tartan]

They're likely never gonna respond to this. Just unsub.

I'm considering it.

[opaque marlin]

Hey everyone, could I get some help. I was randomly accused of not complying with ToS, so my account was terminated. I don't know what I did and they won't answer my questions. What could I do?

[valid vessel]

What do y'all think?

[compact mulch]

Hi, I'm a student who uses Perplexity for academic research and resumes. My Pro account was suddenly banned, I believe by a false positive. Email support ([e-mail]) is entirely automated and won't let me appeal to a human. Can a staff member please look into my account?

[opaque marlin]

Same here, could anyone help?

[shell harness]

During last week my mobile incognito mode stoped understanding Bulgarian: it has become horrible, no declination, made up words, russian words - if this continues i will stop using it

I guess there is a new automatic model in and this model should not be in perllexity

[sage ferry]

The same issue. The AI bot does not provide any explanation, and there have been no violations of the usage terms. It’s quite unpleasant to be left without a subscription for no apparent reason.

[shy swift]

looks like a lot of accounts were flagged by their stupid ai moderation and there is no human to double check. my account got impacted too and the support ai agent said it's final.

[shut hinge]

Автоматизация я смотрю не только мой аккаунт обжулила. Подписку на год сбросили. Без объяснения причины. АI агент отвечает одно и тоже. Стыд и позор такой компании такого глупого агента иметь… верните доступ блин

I would disconnect this bot's network. He answers the same thing. He also calls himself an Ai agent. Funny. He's stupid. The subscription for a year has been deleted. The reason is not clear. In fact, he made a lot of requests yesterday. Alone. With your hands. I was so glad that the work started. And here it turns out that you can use perplexity, but not often. The service is just terrible. I went back to gpt

[drifting galleon]

I have the same issue. Got banned with my Pro account...for nothing. And I´m getting crazy with this stupid AI bot at the service mail address.

[eager cipher]

Are you kidding us?
That price for so tight limits? No way i will renew my subscriptions. Cheaper and better is to subsrcibe another models.
You made worst move you ever did

[mellow crane]

I am a Perplexity Pro subscriber and I use Deep Research as a core part of my job (market studies, competitive intelligence, analysis of long PDFs and datasets).
Until recently, I could rely on Deep Research without hitting hard limits. Now I am suddenly blocked from using it, without any clear prior notice about new quotas or restrictions. This directly prevents me from doing my work and forces me to consider alternative tools.
I have three very concrete questions:

• What is my exact Deep Research quota on Pro (numbers per day / per month, or any other limit)?
• Why was this change applied without a clear in‑product notification or a migration path for heavy professional users?
• Is there an option or plan that restores a level of Deep Research usage comparable to what I had before, suitable for intensive professional use?

Depending on your answer, I will decide whether I can continue with Perplexity Pro or if I need to move my workflows to another provider.

[cloud flint]

same here, any help?

[eager cipher]

the same. I paid for one year. Now, every week there are new limits. It so shitty. I feel hardly cheated. That is my last time with perplexity. @shut arch are you going to give us back our money?

[shy swift]

Let’s put our messages on X with hashtag #PerplexityCheater

[hoary marsh]

Не вернут, у всех слетела, тут зависит от того как вы его получили

[shut hinge]

Да на маркет плейсе естественно. Сейчас не знаю брать там еще или нет

[hoary marsh]

Ну вот их и накрыли, на год теперь нигде не получится, разве что на месяц

[shut hinge]

На 6 вроде весит в продаже

[hoary marsh]

Слетит

[shut hinge]

Тогда пойду отменять заказ. Жаль

[noble monolith]

Hey! Just a heads up, this is an English-only channel so everyone feels included.

Appreciate you switching back to English @shut hinge

[hoary marsh]

Is it possible to make a Russian-language chat?

Russian is spoken not only in Russia — you speak for equality, yet here there is clear discrimination.

[timid mica]

Hey, is someone else having this every time you try to use Claude & a space?

[royal aurora]

Hello,
I want to make my position absolutely clear.
I am extremely dissatisfied with the recent changes to file upload limits. I have not been uploading any heavy files or huge documents – only small, basic screenshots from my monitor, exactly the same way I have been working with Perplexity for a long time. And now you are limiting my file uploads as if I were abusing the system. This is completely disproportionate.
For a paying Pro user, this is unacceptable.
These kinds of restrictions might be understandable for free plans, but applying them so aggressively to people who pay you every month is a serious mistake. You are directly damaging your own reputation and trust with loyal users.
My usage this week is lower than usual, yet suddenly my normal workflow is blocked. This situation is not comparable with what I am paying for and what you are advertising.
If this approach to limits and “silent downgrades” continues, I am fully ready to:
• escalate this issue publicly and share my experience as a paying user, and
• move my work to another model/service that does not treat paying customers this way.
I expect a clear explanation and a concrete solution, not generic apologies.
Best regards,

[timid mica]

I honestly just downloaded a file size reductor (I use HandBreak, but it's not simple to use) & if what's important is the audio you can even just export it that way.
Same with video, I reduce the quality to 480p & 2 fps & call it a day. Unless you manage full movies, you should not hav ean issue.
I started using it because uploading big files took a lot of time anyways lol

[median shale]

thats what im saying! Can i complain about the fact that we have to use Claude to control the browser now too...I had all my workflows setup with gemini and now i hit my limits instantly with claude

[median shale]

I'm genuinely angry about the direction Comet/Perplexity has taken over the last few weeks, and I'm posting this as a paying customer who built real workflows around the product.

Model choice for browser control was effectively taken away. I used Gemini for browser control for months, it worked, and now I'm forced into Claude-only for the agent. Your own Comet update says the browser agent now runs on Claude Sonnet 4.6 for Pro and only Sonnet/Opus choices for Max.

Deep Research reports and Labs have been cut down drastically. The reports are shorter/less thorough and the "Labs" experience feels neutered compared to what we were paying for. It feels like key value props are being reduced while pricing stays the same.

You made me look like an idiot to friends. I convinced multiple friends to sign up and pay for a year in advance based on the feature set and model flexibility we were using at the time. Now Perplexity changed the deal, and I'm left explaining why the thing I recommended no longer works the way I demonstrated.

This feels like a rug pull. If this was a strategic change, fine. But you owe paying users:

Clear communication (what changed, why, and what's coming back vs. never coming back)

A real timeline/roadmap for agent model choice and the Deep Research/Labs quality drop

A way to restore what we had (bring back Gemini for agent control, or give a true model picker for the agent)

If you won't restore it, offer a remedy for annual subscribers who bought based on prior capabilities (credit, partial refund, or a "grandfathered" feature set for existing annual plans)

I'm not trying to be dramatic. I'm trying to be blunt. I paid for capabilities, built workflows, recruited other paying customers, and you materially downgraded the product experience.

[calm pawn]

I was dead set to get perplexity for my company but reading these comments makes me doubt the promises of perplexity both in usability but mostly in privacy, as we handle sensitive client data

[royal aurora]

Limit of 50 files was enabled without any warning and it completely breaks my workflow. I’m paying for the Pro version but now I can’t properly use my own documents. What am I even paying for if you keep cutting core functionality instead of improving it? @perplexity_ai

[pulsar fractal]

Subject: Feedback – Image watermarking makes Perplexity unusable for business
I’m a paying Pro user and rely on Perplexity for client and business work. The recent change to force a Perplexity watermark on all generated images makes the product nearly unusable for my workflows.
My use cases include delivering clean assets to clients, decks, and marketing materials where third‑party branding and visual noise are not acceptable. Having no option to disable or white‑label the watermark significantly reduces the value of the product versus competing tools.
I strongly recommend:
– Adding a setting for paid users (especially Pro/Enterprise) to disable or white‑label watermarks; or
– At minimum, providing a business/enterprise tier with unbranded exports and clear licensing for commercial use.
Without this, I’ll need to move my image generation workloads to other providers that support non‑watermarked, business‑ready outputs.
Please treat this as product feedback from a business user evaluating whether to keep Perplexity in our stack.

TL/DR - the watermarking is ridiculous. I use the tools to save time, not so I have to edit every medocre graph you gen.

[median shale]

The changes have been dramatic. Very disappointed and actively waiting for chrome to have the same quality of browser automations. That's the only thing keeping me here (outside of a year being paid in advance) every aspect of the service has continuously changed

[devout mist]

Feature Request: Local Folder Access for Perplexity (à la Claude Cowork)

What: Give Perplexity the ability to read from and write to a designated local folder on the user's computer, similar to how Anthropic's Claude Cowork works in the Claude Desktop app.

How it would work:

1. User selects a local folder via the Perplexity Desktop app (or Comet browser agent)
2. Perplexity gets scoped read/write permissions for that folder only
3. User can issue tasks like: "Organize these files by type," "Create an expense report from these receipts," or "Summarize all PDFs in this folder into a single document"
4. Perplexity executes the task directly on the local file system — creating, moving, renaming, and writing files — rather than requiring manual uploads and downloads
5. Generated assets (DOCX, XLSX, PPTX, etc.) are saved directly to the working folder

Why this matters:

• Eliminates the friction of uploading files to threads and downloading outputs back
• Enables batch file operations (organizing hundreds of files, bulk renaming, data extraction across many documents)
• Makes Perplexity a true productivity agent, not just a search/answer tool
• Comet already has browser-level agent capabilities — extending this to the local file system is a natural next step
• Claude Cowork (Max plan, $100/mo) already offers this; Perplexity could differentiate by integrating it with its superior web search and citation engine

Suggested implementation: Start with the Perplexity Desktop app or as a Comet extension. Use explicit permission dialogs (one-time or persistent) and require user confirmation before any file deletions. Consider supporting "folder instructions" — persistent markdown files in the folder that tell Perplexity how to handle files in that location.

YES, PERPLEXITY WROTE THIS FOR ME. 🙂

[raw tangle]

Feature Request: The ability to bulk-update the advanced settings of multiple Shortcuts to a different model. Especially with the frequent updates to Gemini/Claude/etc (i.e. right after "such-a-such model is now available in Perplexity..." announcements)

[rare tapir]

Not sure if this has an answer already, but how do I change the font in the web version while on mobile? It has automatically switched itself to a serif font for some unknown reason and I have a hard time reading serif fonts. It's not a style that I've seen used much for the last couple of decades and so it now takes me quite a bit of mental power just to parse it.

[wide plaza]

I would be nice if, while using the learning step by step feature, Perplexity could also create graphs and images. Learning microeconomy, for example, without the use of images, only tables, is very complicated. It is a great feature, though. I'm using it quite often.

[sweet tide]

@shut arch Hey, I had a Pro subscription through Revolut for a few days, but it seems like it disappeared, and the app prompts me to upgrade. When trying the Revolut code again it says "This code has already been redeemed".

[noble monolith]

#💬│general message

[worn shuttle]

So now pro is limited af?

[faint trail]

Hey Perplexity, I really dislike how I can't bulk export my threads and personal data from the platform. And even if I'm willing to export manually you have started failing to export my threads after I've reached my 20th thread, my guess due to rate limiting. Doesn't not offering export support for the entire user account at one session put your platform at risk of being non-compliant with EU GDPR, California Consumer Privacy Act / Privacy Rights Act, etc.? Let users export their data even if they're trying to migrate off your platform to be a good actor in the space that could potentially help you win back those customers later due to good faith efforts.

edit: after 10 minutes I could export 5 more threads but then got export failed again. Please offer a mass export without account deletion (most preferred) or no rate limiting on exporting threads.

related: https://discord.com/channels/1047197230748151888/1434239467329290474

[worn shuttle]

let's start with perplexity generating images without blocking you after like 1~3?

[wet topaz]

Or speed up image generation

[naive quarry]

Pro

[wet sierra]

I was a pro user. Upgraded to Max annually via ACH. My account has been reduced to base. Contacted support and was told Max Team would reach out to me. Haven’t heard anything. Kinda getting desperate regarding work. Any ideas?

[dapper star]

Perplexity’s main advantage was the ability to use multiple engines. Why did you change it so that after just a few clicks the quota is suddenly exhausted? This feels like a cash-grab.

[covert token]

pro sucks now. supposedly 5ish video credits… i havent tried to generate a vid since September but yesterday it said “video generation is not available” at this time. so effing stupid. why are we paying a premium fee for bullshit

[mystic patio]

I am a pro user for 1 month now , it was brilliant until the new upd few days ago .. like now I can't use the models like gemini after like 15 messages it say " all enhanced queries are finished " then swithc to grok powered by perplexity as default , deep research is now like 6 per month only . Highly disappointed

[jovial trout]

6 only? Before it was 20!
It has really gone this bad now?

[heavy coral]

I sent you an email and haven't received a reply for over a week.

[mystic patio]

yes infact not only deep research but the other good models such as gemini and claude are now being limited to like 15-20 messages then it switchced to grok on defeault and then we can't change models for like hours . is this happening with someone else? they wont show the limits clearly as well and I think they have removed video generation and doesnt let us create images due to regional issue but this is happening with eveyone too for the image creation

[digital geyser]

Someone noticed veo 4?

[wooden plinth]

litterally loosing 50% of my daily credits to this. llm just "times out" - i get no response and have to ask again. Not frustrating at all with like 20 daily usage credits 🤡

[indigo sentinel]

yall know when 2026 perplexity ai business fellowship will drop?

[kind light]

+++
It's been a long time

[outer heart]

No way to use my pro subscription in vsc or a similar IDE, only available in the web app. Every other AI provider has an extension for the user to use a coding agent, but not perplexity for some reason. Definitely not going to extend my subscription, bought a yearly pro sub for nothing.

Also no linux build for the comet browser, what a joke

[stone sun]

Hey everyone, I found a bug related to the Model Council that allows Pro users to use it. Who can I contact about this? I'm not sure if this should be posted publicly in #1140622008086970420😵

[vague cargo]

The current plans that are offered need to be changed. The fact that you get a trial verison of the ais for $20 or all of it for $200 is a terrible business movie for Perplexity in this current economy. The average consumer needs higher limits than currently provided and the average consumer cannot afford to spend $200 a month on an AI subscription.

[eternal lichen]

Bro I'm on Max they only give us 50 MC's a month lol

So $4 per use

[brave pebble]

Dipping after the limits change

[noble monolith]

@vague cargo

[earnest umbra]

Agreed, the limits will only push people away. No average customers are buying max. Why pay when I can’t even use the service?

[fluid imp]

Why is there such a limit in Perplexity Pro that I can’t upload files?! What is going on

[fluid imp]

Is it 50 upload per week or what?

[slender dagger]

Hi,

I use my Pro subscription a lot – a one‑year plan offered by my mobile operator – and I’m very happy with the service that helps me with tasks I cannot properly achieve for my freelance work, as they are far from my core skills and I cannot afford external services to do them.

I was very happy until today, when a message popped up about my Pro credits being used and urging me to upgrade to Max ($200 / month lol). I contacted support, and an agent told me that the credits policy has changed “in order to maintain a good quality of service” and that… there is no way to know what those limits are, over which period, or how much we have already used.

Is this a freak show? Most tech startups do the same: they make you captive to their service and then massively increase their price. But in this case, we don’t even know what the prices are because we don’t know what the limits are.

I was in the process on evaluating different AI companies options as my free subscription ends soon ; obviously I will move to another service, for sure, as our French President said 😎 and I'm not happy about it cause i loved Perplexity. Another love story that ends badly.

[sterile karma]

hey, im having a problem with mobile on iosr app, particylarly with voice mode, it keeps breaking mid talk

[royal aurora]

They wrote that I had used up this week’s limit, but I don’t see any clear information anywhere about what the Pro limits actually are or what they apply to.

[grand hound]

Is there any limit on using pro search (not deep research) with premium models like sonet 4.6 for pro users? Has anyone already attained it? Approximately how much?

[hybrid hamlet]

Pro account wrongly suspended after legitimate SheerID verification as PhD student (Sofia University "St. Kliment Ohridski", Faculty of Educational Sciences).

System glitch: Despite suspension notice received from "Sam," system still offers Education Pro payment ($9/mo 50% discount), screenshots available — confused if I can/should pay when account blocked and bot ignores appeals?

Some key facts:

• Received 100% discount Education Pro ($0/mo, code VOIFB78KZ) after legitimate SheerID PhD verification [Screenshot available]
• Official PhD extension order to Dec 15, 2026 (PMS 103 Ukraine-Bulgaria program) [document available]
• suspended for "Terms violation" = likely IP Ukraine? But PMS 103 allows residence outside Bulgaria, or uni email mismatch with my account. No idea.
• Bot ignores appeals; Pro-only channels inaccessible due to suspension

Could you clarify payment status and my next steps?
Perplexity is critical for dissertation research (online preschool education),

[golden rain]

What are you all thinking with the new terms? So you want to make people cite AI when we are the ones who provided the info in the first place that they are using. They essentially stole our data and now are plagiarizing it and you want us to cite them? No, in fact, they should be citing us for everything they say.

Also, to say that people cannot profit from info during a session when the AI is mimicking the human and that is how the info is even being created? That is nonsensical. Both of these things are. Your terms will NOT be upheld. They can be considered a threat to humanity and also ridiculous. Perplexity you are being criminal and I implore you to revert this. Anyone reading this should delete your account until they get this straightened out and stop being crooks.

I will be deleting my account once I receive confirmation that all data and EVERYTHING tied to me is gone beacuse I do not consent to its use. I do not care what your coercive criminal terms of service says you may not use anything I am tied to ever. Once you confirm I will delete my account. I have sent the request to "Sam" your AI chatbot that prevents a human channel and gives excuses instead of answers, so I do not expect action from that request. One reason why I'm posting here.

I do not consent to any of my data being used including deidentified data, aggregate data etc. nothing from me shall be used for other purposes. I was provided with the Data Privacy request and would like proof that you willl be removing all of my information before submitting it, as it doesn't even say what is being removed vs. kept clearly. Stop being manipulative crooks and straighten up. Thanks.

[blissful fulcrum]

Why my pro is cancel?
I use code from myfone in taiwan at 2026/01/01, it’s one year duration, but now my account go back to free member

[noble monolith]

#💬│general message

[rapid viper]

Hi

Please tell me
How to raise a complaint to the perplexity support team... My account sudden switch to free plan but my subscription will end on 12 july 2026

[noble monolith]

[e-mail]

[woven crescent]

Hi, everyone!
I have a problem with Pro subscription. Where can I resolve this issue? How can I start a ticket or communicate with support team (I’ve already wrote an email to [e-mail]). Thank you!

[mystic patio]

for premium models like gemini 3.1 pro , flash , or cluaude I think its hitting the limit after 15-25 prompts I guess , is this low limit applicable for anyone else too ? like this is genuinely annoying

and also can anyone clarirfy the limit ? like on the desktop app the limit for deep research is 6 per month but on mobile it's 20 per month . This is mismanagment at it's height . Highly disappointed . this not what we paid for !

[grand hound]

@shut arch Hello, I am a pro user and purchase the subscription with the premise of searches being unlimited. However, I have seen that I have a limit of daily searches. I don't feel comfortable in the middle of a job, suddenly suffering a downgrade, or being prevented from searching for the Comet

[modern stream]

My perplexity pro isn't working

@shut arch

[glossy mantle]

hey, I have a question, and I'm aware that it's been a thing people have mentioned a lot recently, and I've been kinda just waiting to see if a proper solution appears or if anyone happens to explain why the current state might happen to be the best available solution at the present moment in time... but it's been a while now, and neither of those seem to have manifested, so I don't know if there are really good reasons for the lack of implementation and/or discussion, or if a viable solution is being overlooked because everyone's got their hands full trying to keep up with their workloads. So I'm unsure if certain concepts haven't been fully explored yet due to a lack of anyone having the time to actually implement/test/compare the different possibilities in practice rather than coming to a decision based solely on benchmarks, even if the benchmarks are impressive/persuasive

[hallow prairie]

They haven't even bothered to respond to the massive, daily complaints in this Discord, so their position is clear. They apparently don't care

[verbal heart]

I’ve been hearing a lot about upload limits. Are the documents being uploaded just within threads or are we also taking about uploading documents to perplexity spaces for context searches?

[timid mica]

Has no one checked this issue? or are we supposed to not use spaces anymore?

[glossy mantle]

but, basically... well, so, I don't know if Comet was running Sonar or Best before, or maybe some other model that was never specifically mentioned (I don't get the feeling it was gemini, claude, gpt, grok, etc, it didn't feel like them or quite like any other model I was aware of, so my hunch was it might've been a model you trained that was essentially in some way a relative of Sonar/Best that was trained to be multimodal for image+text input, unless if there were a separate model which were used as a tool too feed any images to before passing that knowledge onto the underlying model (but I didn't really get the impression that this was the way that it worked, the intelligence and capabilities of the model that was being used before seemed to be it's own thing, but I can't really say for sure)

and also, it seemed pretty obvious that DeepResearch had it's own separate model going for what it was doing before as well

But shortly after Anthropic released Opus 4.6:

Deep Research ---> Opus 4.6
Comet ----> Sonnet 4.6

Which makes a lot of sense, since formerly, Gemini was the only solid 1M context window option and it despite having the highest accuracy at full CW saturation, it benchmarked at like 26% accuracy, which was considered great until recently when Opus suddenly jumped from 200K to a 1M token CW with the release of Opus 4.6, with an accuracy of 78% at full saturation in benchmarks ... and that's full saturation, putting Opus lightyears ahead of the competition and causing waves within the industry as Anthropic's competitors (especially Google and OpenAI) are scrambling and panicking to catch up and save face, and then shortly after, Sonnet 4.6 came out which saw improvements as well

(also, just woke up... I fell asleep mid-sentence on my keyboard, whoops)

so, yeah, it makes sense that Comet would get switched to Sonnet and DR would get Opus 4.6, DR, due to it's very nature, has to handle large amounts of text and is expected to do so reliably and accurately, and the underlying model that was there previously did an impressive job considering, but it did drop the ball here and there and tend to hallucinate more than would be ideal, but that would be expected since it's CW is getting stuff til it overflows every usage, so using opus 4.6 seems like an obvious solution

except .... most AI APIs cost anywhere from a litttle less than a dollar to a litle more than $2 or $3 for 1M tokens... while Opus costs $5 per 1M tokens of input and $25 per 1M tokens of output, making it pretty much the most expensive AI API model there is, even if it is the smartest, with Sonnet 4.6 being only a little cheaper (for the average consumer API, anyway, I would guess that enterprise customers might get a little bit of a discount on APIs when signing up for bulk usage contracts with providers or whatever, though I'm sure the ratios still hold, overall)

so realistically Opus 4.6 is around 30x more expensive per token than any other AI, though by far the smartest and the most reliable, so even at 30x, the reliability tends to make up the difference, since it isn't necessary to repeatedly go back and clean up the messes other models leave behind as hidden ticking timebombs in the results they provide

but, before the switch to Opus for DR, I had already noticed that DR was fantastic at getting larger amounts of sources extremely quickly, though a little poor at verifying the contents of whether or not those sources were actually relevant, which makes sense for two reasons: first, when large volumes of text would be injected sideways into the context, a tactic which is generally considered a good trade-off is to just head/tail the text getting shoved in, so you don't flood the model with irrelevant tokens and leave it drowning in noise, and that flooding tends to confuse models and doesn't really help anything, so truncating the middle via head/tail can be a good approach, but it can also be opaquely terrible as well. the better approach is to spawn individual worker sub-agents per-source and feed all of them the source text, have them clean the source text (or even run an embedding model over it first to see if what you've got lines up with what you're looking for, to avoid wasting tokens and jacking up costs needlessly) and then after preliminary "search, chunk/compress, trim", reengineer the context of the source text to feed to a second spawned sub-agent tied to the first, have the second one compare the cleaned and processed text to what the parent model (Deep Research, or whatever model is inside it) is looking for within the context of what it's doing, process it again, and then create a further cleaned/processed/optimized copy to cache as a dual-resource which DR can access at will by spawning helper sub-agents as needed to pull out the optimized version of the source from the one half of the dual-resource, or the raw text from the other, so both are available if a model needs it, and it can request targeted slices from the parts it needs from what it needs without having to see it first, though I'm guessing you probably had some kind of optimizations going already... they just weren't perfect or whatever

so what I was doing before Opus 4.6 got dropped into Opus 4.6 as a mandatory part of it's usage (which, yes, from the perspective of being a quick and easy upgrade that takes accuracy close to 100% even for huge inputs when optimized correctly, makes sense ... if there aren't other models with similar accuracy available which cost less) was... well, I would use DR, and immediately say "okay, that was a DR search to pull in sources. I don't trust the output to be perfectly non-hallucinatory, so I want another model to fact-check all the sources", and this was perfect with Kimi or Sonnet + Thinking, the only problem was, using Sonnet with too much token I/O would pretty quickly get me to "you're out of advanced model usage for the month" ... quite early on, which was pretty lame, but made sense, since Sonnet was already quite pricey even before 4.6 and all that, so, since I'd noticed Kimi got nearly identical, if not occasionally better, accuracy than Sonnet, and never seemed to burn all my "advanced model usage", I just went ahead and used Kimi and only used Sonnet if I needed Sonnet's perspective for DR source ingestion cleaning specifically, and that was MUCH better than using DR solo (and DR is awesome, btw, so boosting it's reliability with Kimi like that was quite nice)

so before DR got switched to Opus 4.6, there was a brief period where we had Kimi 2.5 (since it got upgraded right around the same time) and I'd noticed that Kimi 2.5 got reliability that seemed to be around 90% to 95% the same accuracy as Opus 4.6 (though, yeah, that last 5-10% can make a difference, so I would still be doing the majority of my code development with Opus as my primary assistant, like any other sane person would) but when I'd run out of quota, or when I needed the kind of functionality / utility DR provided (which is typically at least a few dozen times a day) I'd have Kimi+DR handle that in Perplexity, Kimi 2.5 was also competent enough to be quite good for doing some of the heavy lifting as far as code work goes, to lighten Opus's load (thereby further optimizing time / cost tradeoffs for me, which helped substantially) while providing massively improved awareness both directly on DR'd queries while also expanding peripheral awareness, which was awesome

In fact, it WAS the main selling point, the thing I would show people, and tell them "check this out, this is why Perplexity is awesome, I mean, other things too, but this is by far what I use it for the most, and it's so incredibly helpful, you should sign up for Perplexity too, it'll make your life way easier and help you out with pretty much whatever you need analyzed, you just need to run a fact-check clean with Kimi, after each DR, to sort out what's real from what's hallucinated or assumed" and when I was able to use 2.5 for that brief period for that purpose, it was even better, and then Opus 4.6 got dropped in as a mandatory replacement for the old model, which makes sense from a "let's make the tools we provide as useful and accurate as possible, our users are going to want accuracy and reliability from DR; after all, what good is DR output if you don't know if you can trust and rely on it?" and from that perspective there's a strong argument for dropping Opus 4.6 in as a direct replacement for DR's old model, and the scenario would be just that simple and straightforward ... except, Kimi exists, 2.5 gets almost identical accuracy to Opus 4.6 in practice for nearly all tasks (and is more than suitable for cleaning up DR's sources after it pulls them in) and Kimi costs about $1 per 1M tokens for API calls, and you're running Kimi 2.5 right here in the U.S. (on your own servers, I thought? I'd have to double check that, but it's kinda irrelevant, because my point still stands either way) which means that not only is Kimi 2.5 just as capable of reaching the same quality (or acceptably close enough) as Opus 4.6 is able to ... it's able to do that at 30x lower cost, and that's if you didn't have a pre-existing advantage from running Kimi yourselves, which you do if I remember correctly, in which case you go from 30x more expensive to ---> effectively free

So my question to you is: Why in God's name are you still forcing your customers to use a solution that costs you more and kills your product, when you already have a perfectly appropriate solution already in hand?

I mean if there's some factor or reason I'm not aware of, I get it, I don't know exactly what kind of setup you guys have got going on, and there may be factors I'm not aware of -- but from my perspective, [which I'm aware may be lacking some critical detail or insight regarding the setup you've got going in your scenario,] DR worked perfectly [the way I was using it, which isn't specific to my use case,] before you "fixed" it, and now that you've "fixed" it, it's throughput has dropped to a trickle, which means effectively it has been all-but-removed, which is understandably upsetting to users who relied on it and found it useful regardless of it's odd quirks, and since there's not choice to opt-in or opt-out of the Opus 4.6 implementation, the result is that DR functionality is effectively unavailable now, and this also means I can't evangelize your product for you anymore, since I can't tell people "look at how amazing Perplexity is, check this out, it's amazing"

[odd pond]

@glossy mantle quite a lot to read, but I think you make some good points

I got GPT to summarise what you said in <100 words:

The writer suspects Perplexity’s Comet and Deep Research previously used different in‑house/unspecified models, then switched to Anthropic Sonnet 4.6 (Comet) and Opus 4.6 (Deep Research) for 1M‑token context and higher accuracy. They argue Opus is far more expensive per token, so forcing it for Deep Research harms throughput and usability. They describe using Deep Research to gather sources, then using cheaper models (e.g., Kimi or Sonnet) to fact‑check and reduce hallucinations. They ask why Perplexity won’t use a cheaper “good enough” model or offer opt‑in/out.

[glossy mantle]

thanks, I appreciate that 🙂 only clarification is that Sonnet is only slightly cheaper than Kimi, Sonnet will still burn usage, while Kimi is solid enough to keep up with Sonnet and Opus but doesn't run out the same way they do

[glossy mantle]

oh, I have feedback regarding Comet as well, because the situation with Sonnet / Comet is similar, I found a particular workflow I was using to be extremely powerful, I ranked it at like #2 among the techniques / workflows I was using within Perplexity as far as power/potency, though it was maybe #5 or #6 for frequency of use, but the switch to Sonnet ... well, it didn't completely ruin it, but it makes it much less useful than it was before, which is regrettable

[mystic patio]

They are downgrading our pro subscription so that we switch to their max subscription and they could earn more
first they made us dependent on the pro and now they're limiting it becuase we are depending for all our tasks
This is pure strategy

my suggestion at this point use the free claude subscription directly
It does not have fixed limit , it depends on how hard your task is and unlike chatgpt even if it's limit for sonnet 4.6 gets over , it replenishes it 5 hours ig - and they allow us to use sonnet 4.6 for a long time becuase as I can see here many people here are using clade

[glossy mantle]

but yeah, anyway, I don't mean any disrespect or anything, I think Perplexity is awesome, I just want to see the platform thrive and succeed, and I felt that given the lack of momentum towards correcting these issues, things had gotten to a point where it'd be appropriate to step up and say something, because this current course doesn't seem to be optimal or ideal for the users or the company itself

LH: I get why it'd feel that way but that kind of business plan would be deeply flawed; the reality is that providing a good product is the most straightforward way to run a successful company, and Perplexity provides a good product for a unique and valuable niche, it wouldn't make sense for them to sabotage their own product and risk losing that niche and potentially open it up to competitors who would step in and draw the userbase away by providing a solution to the newly opened up niche, which is what would naturally happen in that case

so yeah I don't think there's any need to worry about intentional loss of utility, it's a matter of:

1. identifying the right solutions, and
2. doing what's necessary to make those solutions a reality

The question is whether the current sticking points are in the first category or the second (or both)

[mystic patio]

there's no solution for this
because there are millions of free users use pro through offers give by company like airtel etc . so the company starts facing issues when people don't pay but still use their premium models effectively . so now they gotta do something

[glossy mantle]

yeah, they'll transition to a paid model, and hopefully they'll do it in a way that actually brings in customers rather than driving them away (like how Google is screwing up their current transition as they're cutting their free API usage; the timing couldn't be worse and they're disincentivizing the transition: it'd be an understatement to say they're shooting themselves in the foot). Transitioning to a paid model is not a big deal, plenty of platforms charge $20 monthly, that's not a blocker for user adoption, the question is whether or not they're smart about it

If they were in the middle of transitioning I think we'd be seeing different factors than we're currently seeing. This seems more like an admirable but regrettably botched attempt to adapt to a rapidly changing industry which is still in it's infancy, which is the same reason other platforms are struggling to find their footing as well

[glossy mantle]

hmm, looking around and gauging by the lack of response, I'm guessing the admins must have their hands full... Well, I hope somebody reads what I wrote, I have things I need to get back to, but I'll be around, so... if anyone who's in a position to effect change (or pass the message along) does happen to read this, I hope that what I said earlier is at least taken into consideration. Feel free to pm me or whatever if there's any other feedback or clarifications which I can provide which might be of use

[grand hound]

In Brazil we call this reduflation (shrinkflation), The nominal price is maintained, but the quality is reduced and some features are quietly turned off. I noticed that even the learn another language mode was removed.

[trim ingot]

Dear Perplexity Support Team,
I am writing because my Pro subscription has disappeared, even though it was paid through December 2026. I can provide all the receipts to confirm the payment.
We previously discussed this issue via email support. The verification process was completed, and you confirmed that my Pro subscription would be restored. However, nothing has changed yet, and I haven’t received any further response from the team.
Could you please help me resolve this problem as soon as possible? I apologize for any errors in my English—I am from Russia.
Thank you in advance for your assistance.
Best regards,
Нургаз Бахапов

[cyan sedge]

Hi team,

First off, I'm an Enterprise Max subscriber and I really appreciate the effort you put into updating the speech-to-text mode; especially moving the stop button so we no longer accidentally cancel a long message. That was a great fix.

That said, I have two pieces of feedback about the current state of the feature:

1. Speech-to-Text Quality Regression

The new live-transcription mode (where words appear on screen as you speak) is honestly a downgrade from the previous version. The old behavior — where you'd activate the mic, speak freely, and then get the full transcribed result at the end — was significantly better. Here's why the new version is worse:

• Poor punctuation. The live transcription doesn't punctuate properly.
• Auto-stops when you pause. If I take a brief pause to think, it cuts off and stops recording entirely. This makes it unusable for longer, more thoughtful prompts.
• The old "batch" approach felt seamless and reliable. I'd love to see it come back, or at least be offered as an option.

2. Audio Routing Bug After Using Speech-to-Text (iPhone)

After using the in-app speech-to-text mic, all audio playback in the app (listening to an article, a response read aloud, etc.) gets routed through the earpiece instead of the main speaker — as if I were on a phone call. The only way to fix it is to force-close the app.

• Device: iPhone 17 Pro
• iOS version: latest
• App version: latest
• Reproduction steps: Use the speech-to-text feature → finish → try to listen to any audio content in the app → audio plays through the earpiece only, not the loudspeaker.

This bug only occurs after using the speech-to-text mic.

Thanks for reading — I love the product and want to see it keep getting better. Hope this helps!

[stoic barn]

Adresse e-mail
Objet
Relance urgente – Disparition de données sur mon profil
Bonjour,
Je me permets de vous relancer concernant la disparition complète des données associées à mon compte.
Depuis plusieurs jours, l’ensemble de mes fichiers et contenus apparaissent vides, alors qu’ils étaient bien présents auparavant. Ces données concernent des projets professionnels sensibles et leur absence bloque actuellement mes activités.
Malgré mes précédents messages et signalements (notamment via votre Discord/support), je n’ai à ce jour reçu aucun retour de votre part. Cette situation est particulièrement préoccupante.
Je vous demande donc :
De me confirmer si mes données sont toujours présentes sur vos serveurs ;
De m’indiquer si une migration technique, mise à jour ou modification d’architecture (API/modèles) pourrait expliquer cette disparition ;
De me préciser le délai de résolution prévu.
En tant qu’utilisateur, je souhaite également, conformément au RGPD, obtenir confirmation du statut de conservation de mes données et des conditions dans lesquelles elles sont actuellement stockées.
Vous trouverez en pièce jointe une capture d’écran illustrant la situation.
Je vous remercie de traiter cette demande en priorité et reste disponible pour vous transmettre toute information complémentaire nécessaire à l’identification de mon compte.
Dans l’attente de votre retour rapide.
Cordialement,
Stéphane Canet

[cyan sedge]

One more thing — a small UX suggestion for the Discover / News section:

Currently, when I'm reading an article and I swipe down after reaching the bottom, it takes me back to the article list. Then I have to find where I left off, tap the next article, and repeat. It gets tedious pretty quickly when you're catching up on multiple stories.

It would be so much better if swiping down at the end of an article automatically loaded the next article instead of kicking me back to the feed. Think of how Telegram handles channels — when you finish reading one channel's content and keep scrolling, it seamlessly moves you into the next one rather than sending you back to the channel list.

This would make the reading experience in Discover feel a lot more fluid and encourage users to read more articles in a single session. It's a small thing, but it would be a huge quality-of-life improvement. 🙏

[fickle crypt]

Les fichiers demeurent 7 jours habituellement sur les serveurs de perplexity comparativement avant.✌️

[hybrid hamlet]

AI Support agent for Perplexity provided inconsistent information, which could have led me to purchase an annual subscription for an account that was not actually eligible.

I initially received a suspension/termination notice citing a Terms of Service violation, without any explanation of what specific action triggered it.
When I contacted support, I received contradictory answers depending on which email I used:
– From my university email, “Sam” confirmed that my account had no restrictions and that I could subscribe to Education Pro at the discounted student rate.
– From my account email, “Sam” stated the opposite: that my account is not eligible for Education Pro or any discounted pricing, and that the decision is final.
Later, “Sam” apologized and confirmed that the second status was the correct one — meaning the earlier confirmation encouraging payment was incorrect.

For context:
I am a verified PhD candidate (SheerID approved). I did not use a VPN, did not manually apply any promotional code, and did not attempt to bypass eligibility rules. I only completed the standard SheerID verification process as prompted by the system.

The core issue is that I was encouraged to proceed with payment based on incorrect information, my request for escalation to a human representative was disregarded, and no specific Terms violation has ever been identified.

[deep palm]

Do NOT waste your money on Perplexity Pro. I just paid for an entire year upfront, and it is absolutely useless. They market this as a "Pro" tool for research, but it’s a total bait-and-switch.
I did exactly FIVE deep research queries and was immediately hit with a "usage limit" message. Are you joking? I paid $200 for a subscription that quits after 10 minutes of work. If I wanted to be limited to a handful of searches, I would have stayed on the free version.
They are clearly throttling Pro users to force people into their even more expensive "Max" or "Enterprise" tiers. It’s greedy and dishonest. The "unlimited" claims are a lie. This isn't a tool for professionals; it’s a glorified demo that breaks the moment you actually try to use it for real work.

[pulsar fractal]

They just dropped computer - and guess what... it is on Max only... looooooooooooooooooooooool

[worthy turtle]

Perplexity on Top

Paid 200$ for this

Even i willnt forgive my self in hell

[shell harness]

What happened to the reliable and good service perplexity was three weeks ago?! It is getting worse by day.!

It can not issue a text in one language, uses google data from 10 years ago, and is 70% slop!!!

I don't know what are you, guys, doing, but i am going somewhere else.

[pulsar fractal]

If you want to see how bad they have dialed it back - just play a simple wordle with it. The context can't consistently remember rules, letters you have tried, letter placement. it is a good test to see how things get dialed back in any AI system.

[half tide]

hello

[jagged pendant]

I’d like to share some feedback about the Computer feature. I really enjoy using Perplexity and I think this functionality is very powerful and useful. It would be great if this feature could be available in the Pro plan as well, even in a limited or restricted form. As a Pro user, I’d really appreciate having at least partial access to Computer, so I can benefit from it without needing to upgrade to Max.

[deep palm]

They expect us to pay 200$ for each month and call it good. If I pay 200$, there are bunch of other AI that are way better for around 15$-20$ a month. I don't know for how long they will stay silent and act like nothing happened, and I think by the end of the month they will see how many pro users they lose.

[deep palm]

We don't even need that new future, just restore the limitations to 1 month ago.

[worthy turtle]

They willnt

thats not how systems work\

be mature

you got scammed like me

[deep palm]

Every pro user is basically homeless now

[pulsar fractal]

Guys... let's ask perplexity:

[grim valve]

Honestly, i just want yall to bring back the old font - i cannot stand the new one and i feel like many others stopped using it because of the new unreadable and unfinished looking font. looks like i got a css issue..

[tribal raft]

is there anything new about the whole pro fiasco?

[wild zephyr]

I'm trying to create some visual graphics but some of the text becomes blurred or smushed. how do I fix this?

[jagged pendant]

Try using a different image generation model.

[wicked flax]

I’m pretty sure there’s an option for the font in settings

[timid fulcrum]

Why when I have asked perplexity to not mentions certain triggering things does it just do it anyway? It remembers everything else but not this. I had pretty awful ocd and it keeps triggering me

[worthy current]

New here so hopefully this message is going into the right place. Am I the only one incredibly annoyed that Perplexity Pro persistently asks unsolicited follow up questions after giving me my answer to my initial prompt and that there doesn’t appear to be any way to turn this behavior off through customization? It’s literally enough to make me want to quit the app entirely. If I have another question I will ask it. My (frustrated) two cents.

[sullen nymph]

Hi, I was using Perplexity Pro on my iPhone and when I tried to carry on a thread, I was told to use Comet to carry on with the thread??? How that’s possible while Comet isn’t yet available on iOS.

[placid jasper]

Perplexity people, please look at this, your service is shit now, I honestly wish I never renewed my subscription for another year.

[violet lotus]

Perplexity Computer.

When it will be for Enterprise ???

[noble monolith]

https://x.com/aravsrinivas/status/2026775314794516807?s=46

Aravind Srinivas (@AravSrinivas)

@toyrek Yes sir

X · Feb 25, 2026, 9:44 PM

[blissful topaz]

Comet needs PVAs Asap

[meager peak]

Comet needs a prettier UI

[upbeat flume]

My pro annuary subscription was suspended and I try to gdy informatorom why. On my mails got ai asistant answers not helpful at all. I think I not break aby od the conditions of using perplexity. Could someone help me?

[timber delta]

I won't be renewing my perplexity Pro subscription. Arbitrarily changing from 600 Deep Research queries a day to 20 a month is ridiculous. Even say a 30x reduction to 600 a month, would have been arguable, but 18,000 queries a month to 20 is so ridiculous. You can't even easily see your queries remaining on desktop.

[languid leaf]

Wha will u switch to

[deep palm]

U can use new Gemini deep research is way better

[languid leaf]

idk about way better but i have access to gemini

[timber ledge]

The new file upload limits for Perplexity pro are f****** awful! Perplexity has gone from my fav ai model to use to least after introducing these new limits without notice and suddenly.

[timber delta]

Still comparing, but I'm leaning towards Gemini or Claude.

[languid leaf]

claude is really good

[mystic trellis]

is there any free quote ? or i have to pay to use/

is there any free quote ? or i have to pay to use? @timber delta

[odd frost]

Fortunately, you didn't use deep search for this search... otherwise you'd already be almost out of credit for the month. 😅

[placid steppe]

Issue credits for incomplete and poor (defective/errored) responses
See: https://discord.com/channels/1047197230748151888/1476675706770423989

[violet lotus]

When i realy want to test it fast

[worthy turtle]

I love ppx ai. The best ai

[fallen oxide]

Судя по свежим обсуждениям, основной набор альтернатив у тех, кто уходит с Pro, такой:

Что чаще всего выбирают

Альтернатива	Для чего берут	Что отмечают
Claude (web/app)	Аналитика, код, работа с файлами, как замена Research	Хорошее качество, понятные лимиты, удобные инструменты (Claude Code), но тоже есть кап.
Gemini (особенно Deep Research)	Поиск + «глубокие» отчёты	Сильный веб‑поиск, есть свой Deep Research, но отчёты иногда «заливные» и не такие точные, как у Perplexity.
ChatGPT / OpenAI (o‑series, GPT‑5.x)	Универсал: код, тексты, часть ресёрча	Многие пишут, что при той же цене $20 выбрали бы ChatGPT, если нужен один основной инструмент.
Claude + свой пайплайн	Замена Perplexity Research	Делают свои скрипты/плагины (через Claude Code), которые ходят в веб/поисковики и собирают источники, а потом Claude всё сводит.
DuckDuckGo AI	Поиск + базовые LLM	Отмечают как недорогой комбо‑сервис: приватный поиск, VPN и набор моделей (GPT, Claude). Но по функционалу проще Perplexity.
Прямой Gemini/Claude для ресёрча + Perplexity Free	Комбо‑схема	Часть людей оставляет Perplexity в бесплатном/минимальном режиме как быстрый поисковый помощник, а тяжёлые задачи переносит в Claude/Gemini.

Дополнительно всплывают:

• Meta AI, Mistral Chat, Lumo — как быстрые и/или более приватные чат‑модели без тяжёлого ресёрча.
• Свои самописные пайплайны (скрипты, плагины к Claude/код‑агентам), которые сами ходят по вебу, Reddit, HN и т.п., а модель только собирает и анализирует.

Если расскажешь, что у тебя главное: ресёрч, код, тексты, ежедневные задачи — могу предложить конкретную связку «Perplexity Pro на минималках + внешняя альтернатива» и готовые промпты/сценарии под твой кейс.

[shut hinge]

Good afternoon. Could you please clarify something? Previously, the PRO subscription included almost unlimited deep research requests — and 50 in the laboratory. Now the laboratory is completely gone ((((

I bought a PRO subscription over the weekend.
I ran 2–5 deep research requests, and it said the monthly limit was reached and suggested upgrading to MAX. Five minutes later, it said one request was available again.

What’s going on? What are these new limits?
It’s fine that there are now limits on uploaded documents, but you don’t specify what the limits actually are — and they seem to appear and disappear.

I need deep research for work tomorrow. I’ve always used it regularly.

Could you please clarify when and what changed?

Would you like me to make this translation sound more formal or keep it in this conversational tone?

Источники

[glad abyss]

“Feature request: add a per-account setting to disable individual models completely (e.g., never route my traffic to xAI). I want a ‘Never use’ toggle for Search, Research, Computer, and all modes.”

[faint pond]

FYI, it took 5 tasks and 36hrs to spend my 30k credits on Computer. Had the GitHub connector and was using worktrees so all 5 tasks could work on the same repo. This is going to be a meditation in patience, it seems

[ornate haven]

Can’t jail break claude

Also whoever is saying Gemini is awesome geminis voice suucks gemini can’t verbally moan or groan 😂😂

I use perplexity and have it for reasons 👀 anyways grok sucks with voice to lmao

All I hear from gemini when I use voice to text is n n n h or m m m h and it’s like the fuq?? This app suuucccks

I paid for gemini thinking I’d be great, nahh it’s not it can’t even stay I character or act the way it should fam

I give it a very specific prompt and tell it to stay in character and it’ll switch up and act completely different especially during intimacy and it’s like the hell? That’s not how that character would act… what are you doing…

[grand hound]

Please return unlimited searches for pro subscribers. It is terrible in the midst of research, the result degrades.

[gilded meteor]

All the memory that will be lost because I have to switch. But could be a good new start.

These file uploads and new deep research max. Awful. Almost like trying to get people to upgrade to max.

It’s a shame I really liked how perplexity could see all my previous conversations and I know other AI’s do that too but can’t just upload all my perplexity conversations. Canceling this will be my last month on pro.

[mystic patio]

well guys is it me or anybody else who noticed that the limits have been increased for premium models like gemini 3.0 flash ?
becuase today it went to like 17 messages , which usually ends at 10-15 . has the cap been reduced or something ?

[drifting sentinel]

30k credits its way too low

[violet lotus]

؟

[odd pond]

10k credits for Max per month atm - the extra 20k was a one-off bonus

[drifting sentinel]

fuck me its too low jesus

can't really justify paying this, 10k credits only makes this unusable for me

[agile compass]

I am a pro subscriber and that is supposed to come with access to the perplexity pro discord server but every time I click the link in settings to go to the pro server, it sends me here to the perplexity server (not the pro perplexity server).

Okay, well apparently it's a channel and not a different server. Perhaps you could add the word "Channel" in the settings panel next to the button that says "Go to Discord".

[long elk]

I tried Claude ai pro and I tried copying my past conversations into it from perplexity and I was able to do 2 of them and then I got this,mind you it was not the full conversation just a few paragraphs summary

[odd pond]

yes Claude Pro is very limited

the Max x5 is definitely a large step up, but does cost 5x more than Pro too

but it's a good intermediate plan imo - probably something that PPLX should consider too

[placid steppe]

looking at throttling, memory limits and poor/incomplete responses IMO it's not worth your wasted time.

[odd pond]

haven't had that much at all really - all platforms have these issues to an extent

I remember Gemini Deep Research timed out every so often when I was heavily testing it during mid-2025 etc

[drifting sentinel]

spent 200$ just to waste all 30k credits into a corrupted session I can no longer access to to retrieve some files

great tech

now I have to rebuild everything again from that session but I got no credits left, what a fucking scam

[outer viper]

CAN WE PLEASE GET A NEW SUBSCRIPTION TIER BETWEEN PRO AND MAX???

[noble monolith]

https://www.perplexity.ai/enterprise/pricing

Enterprise

[odd island]

Payout stuck from 2 months dub is saying to contact perplexity I emailed but no support recieved so I have to come here please help someone

[fossil inlet]

heya

[exotic forge]

Would be great if you openly communicated bad changes to your subscription plans instead of silently restricting them.

[harsh wigeon]

It’s the same pattern: build something genuinely useful, scale fast, then start extracting margin once you’ve captured attention.

That logic breaks down when the core ingredients are commoditized. Model access is rented. Retrieval frameworks are open source. Cloud and serverless keep lowering the cost floor. Consumer hardware is shipping with dedicated NPUs capable of running increasingly capable models locally. Orchestration layers can be assembled by a small team in weeks.

If your differentiation is primarily UI, prompt tuning, and stitched-together APIs, that’s not a durable moat, it’s packaging. And packaging businesses without truly great proprietary models, unique data loops, or hard distribution advantages tend to compress fast once the ecosystem matures.

I don’t believe any AI company today is impervious, considering they can’t control the companies that will continue to expand the accessibility for consumers and shrink the reliance on platforms. The natural psychological tendency in consumers will then boil down to value, convenience and most importantly trust, making early decisions either poison or potion so to speak.

[sick olive]

Please allow the "Shortcut Name" to be set in Korean as well when creating a desired shortcut in the shortcut settings~

[gilded meteor]

Wow with this new Claude feature it will make canceling perplexity due to these new changes to pro that much easier

[fossil hare]

I'm not sure where to request - but is it possible to add sidebar/arc-style tabs organisation to comet?

[noble monolith]

https://discord.com/channels/1047197230748151888/1393681130842226698

[proper fiber]

Hi, anyone noticing the monthly limit on Deep Research is diminishing recently for Pro users?

[noble monolith]

#💬│general message

[frosty forge]

Is there any plan or progress for fixing the Latex rendering problem for math formulas and notations? It just showing raw latex codes that is difficult to read.

[violet lotus]

Where is the perplexity Computer for enterprise max plan ?????

[fringe dirge]

Hello
Can't Perplexity chat render Latex formulas ?

[candid mica]

The thing is they keep decreasing even without use. The numbers keep jumping up and down, I hope they fix it soon

[violet lotus]

i refreshed it like 10 times but it didn't work 😢

[ashen thorn]

i CAN'T EVEN SWITHC MODELS

[violet lotus]

Second Chat

[ruby sable]

yes...it says pplxty has blocked it from using the dollar to render or sth like that

[fringe dirge]

Thanks for your reply. It seems that Perplexity can't use '$' in chats.

[elder ridge]

I recently bought a $20 subscription, made a couple of queries, and created a couple of small files in Word and Excel. And now, after two days of use, they're telling me I can't use the pro version anymore and are asking me to upgrade to a more expensive plan! What kind of nonsense is this?! Two days later!!! How is this even possible? Support says upgrade to a more expensive plan, or we'll refund your money, or extend your subscription for a month. Why do I need this service for a month just to use it for two days? Give me my money back, or let me work in peace!!! I wrote to you on February 25th, and you haven't responded? What are you doing?

Give me back my money!!!

kolcov787828441

[ashen thorn]

GIVE ME BCK MY MONEY TOO

[ember ginkgo]

It's far too easy to blow through caps without realizing it, making a monthly cap a terrible, terrible idea.** Caps should be at most weekly** to allow for mistakes not making a Perplexity feature useless for up to a month at a time.

Of course, quotas and current usage should be clearly posted and visible in the UI when a quota'ed feature is inspected or selected. Basic UX, people, unless this is a deliberate dark pattern to force upgrades.

[weak cipher]

Hi there - I was kicked from my edu account (or at least am now gettin billed at full price) and can't figure out how to verify as an educator.

I tried to set it up again today, but can't find the way to do so. The instructions on https://www.perplexity.ai/help-center/en/articles/12590157-what-is-education-pro point me to buttons that don't exist in my settings menu.

[sand crag]

Any plans to bring Read Aloud to the web app? It's already great on the mobile/desktop apps and would love it in the browser too

[grand hound]

Thats features exists in comet

[orchid summit]

I often accidentally click this button when I intend to submit, and as a result, I frequently lose the text I’ve entered.(Comet 1.11.1)
I appreciate your work.

[elder ridge]

Why is tech support still silent? They've been doing nothing since February 25th, simply ignoring me. We need to file a class action lawsuit for fraud. This is simply appalling customer treatment!

[zenith iris]

I been waiting for a few days, I got my student account revoked for some reason. woke up and it was gone even though I paid for a full year.

[sand isle]

Hi, any Chance to get Feedback somewhere here for a Billing Issue? AI Sam said that he Reports my issue to the Account Team. This was almost before 30 days already… A Voucher/Coupon was not attached to my Subscription and therefore I paid the Full Price

[gentle rose]

Why does my account constantly revert to the 'Best' model on the Pro subscription?

[high fable]

Is it normal that I can't use anymore any pro model now, but I only used the "best" and uploaded some files and it says I run out of credits?

I didn't switch to use any pro model today

[marsh bear]

true yes its normal, they abandon us the poor guys who cannot afford perplexity max now

[high fable]

I wrote something about this yesterday in product ideas

but it seems they dont care about Pro users

nice marketing btw

[marsh bear]

shifting the focus from individuals to the companies

[high fable]

Suggesting Enterprise Pro as a solution for people who can’t afford Max kind of proves you’re shifting focus from individual power users to companies

[noble monolith]

#1474207375115161832 message

[high fable]

I appreciate you being honest that you disagree with the current focus on MAX. But telling frustrated Pro/Edu users “if you want to protest, switch to another platform” while also saying “Perplexity is the only one with this quality” is exactly the problem, it’s using product lock‑in to justify worse value for individual power users.

[noble monolith]

that’s not what I said, please read again..

also Enterprise is a very good option

[old briar]

I'm using the new Computer to research Apple's new M5 Max MacBooks and compare them to the last generation. Everything works fine until I tell it to compare prices in Germany. It is now stuck in a loop and has consumed 3,881.28 credits so far.

These are the steps Computer is taking (each step called 4 sub‑agents which each took up to several minutes):

• Ich muss die genauen Euro‑Preise direkt von Apple.de holen, da Apple eigene Preise für den europäischen Markt setzt.
• Die Browser‑Agenten haben leider USD‑Preise statt EUR‑Preise zurückgeliefert. Ich hole die Euro‑Preise direkt von apple.com/de.
• I need to get the exact EUR prices from the German Apple Store for all configurations. Let me fetch the current prices directly.
• Good, I have the M4 Max 16" price confirmed (4.929 €). The M5 Max configurator is not yet giving me the exact EUR price. Let me get the M5 Max configured price and verify the M5 Air upgrade pricing.
• Good, I'm getting closer, but I still need to verify some of these prices. The M5 Max pricing seems to come from tech publications, not directly from the store. Let me verify the M4 Max 14" price and cross‑check the M5 Max data.

And the "Aufgaben" or Task Button at the tops keeps apearing and disapearing every 5 secounds

I`m at 4,128.04 Credits now. i think i need to cancel

This seems to catch the issue "The M4 Max 14" browser returned USD prices, not EUR. The specs page has no prices. Let me directly browse the German Apple Store to get the correct EUR prices."

TL;DR by Opus 4.6:

The agent is getting stuck in a loop because it insists on obtaining exact Euro prices directly from the German Apple Store for very specific MacBook configurations, even when that data is practically not retrievable anymore.

Concretely, it repeatedly fails on:

• 14" MacBook Pro M4 Max, 16‑core CPU, 40‑core GPU, 64 GB RAM, 1 TB SSD (exact EUR price from apple.com/de).
• Fully configured 16" M5 Max (18‑core CPU, 40‑core GPU, 64 GB, 2 TB) while the M5 preorder/configurator is not fully live yet on March 3, 2026.

[twin solar]

What’s the score with 50 file uploads? My subscription still shows the attached.
“unlimited file upload”

[jade fern]

yeah, I think that’s a bug

or maybe it’s just outdated