community-help

[charred parcel]

nobody here can help you with such things, you need to use https://help.openai.com

[cerulean stirrup]

I've tried this multiple times and its not working. Do you work for Chat GPT?

[charred parcel]

no, people who work for OpenAI have gold coloured names

[cerulean stirrup]

CAN SOMEONE FROM CHAT GPT PLEASE HELP: Support and Privacy are both not responding to me. Chat GPT has changed my plan multiple times without consulting me, overcharged me, said they will refund me and haven't, then keep ignoring me when I contact support.

[charred parcel]

nobody here can help you with such things, you need to use https://help.openai.com/

[cerulean stirrup]

As I've said. I have used the Help Center many times and I am not getting a response.
I need an OpenAI staff reply only please. Thank you

CAN SOMEONE FROM CHAT GPT PLEASE HELP: Support and Privacy are both not responding to me. Chat GPT has changed my plan multiple times without consulting me, overcharged me, said they will refund me and haven't, then keep ignoring me when I contact support. This has been going on for 4 months.

[woven tree]

Guys, I have ChatGPT Plus version but my Sora account is locked to 30 generations per day. On the internet it says that ChatGPT Plus versions are the same as Pro version for Sora - meaning I'd have option to generate more vids and also without watermark?

[charred parcel]

Plus users only get 30 generations per day. Pro is required for no watermark.

[woven tree]

Ok. So if I buy ChatGPT pro version, my problem is solved?

[bitter reef]

i thought if you get 10 votes in the daily image channel, your image goes into the hall of fame?

[split hazel]

One star comes from Starboard, so if you see 10 total reactions, that's 9 other people.

[onyx kraken]

You might prefer absolutely no answer then, and I apologize if that's the case, but we are almost 100% community members here. The extremely few of us who work for OpenAI have gold names, #server-staff .

There are the #1070006915414900886 and #1070006151938314300 channels, were sometimes we see gold-named folks discussing, but even the mods here are volunteer community members, and it seems likely that OpenAI does not monitor these channels where we talk to each other, you're not talking to OpenAI at all here.

The only support I am aware of is through the website you already say you are using.

You mentioned problems with charging - If I was in that situation that you describe, I would reach out to my financial institution that I do business with OpenAI with and inform them that I don't choose to continue to pay and am having issues with canceling through OpenAI. Your financial institution should be able to help you regardless of any other situation going on.

I wish you the best!

[bitter reef]

Ohh, so close 🙂 Thanks for letting me know.

[fierce zinc]

Voice Mode stopped working for me on iOS 26.1. Now it won’t work anywhere. Any suggestions?

[kindred oasis]

You're such a kind and lovely creature, Eskcanta. I truly hope that you never harden because of life's adversities or because of other people.

[muted solstice]

?

[static saddle]

Is there anyway we can fight this change in model last night from image generation in chatgpt/sora1? I make a anime series on youtube with them and been getting consist characters for months until last night...Thought i was messing up somewhere until i tried all my old prompts and the artstyles are completely different all around and the image quality isn't consist and not following prompts correctly.

The anime model now feels like generic chinese anime style trying to fake high quality image honestly.

[paper timber]

that's a good way to put it. sadly rn no way to revert

[old lodge]

anyone knows wgats happening? its taking up to 3 minutes for a detailed answer and 1 min for basic! I have the paid version

[hollow lark]

SOMEONE NEEDS TO FIX ChatGPT’s BRAIN

IT DOESNT REMEMBER MY CHATS

[static saddle]

does chatgpt have a api of the old model for image generation? The new model update is ruined for making anime and is slop now. I tried confyui before but it was too confusing setting it up. Any recommendations on a simple set up api with consist anime styles like the old model?

[kindred oasis]

Apparently they're more concerned about beating Gemini's image generation instead of fixing what's obviously wrong with the new model

[stable grove]

Why is openai very lazy on working on the windows chatgpt app?

It's just an electron app💀 easiest to develop.

Why does it have no multiple tabs feature?? This is outrageous.

[sweet crystal]

Hi everyone

I am using ChatGPT to study math, and it explains the lessons to me in Arabic. Because Arabic is written from right to left, I noticed that some math expressions sometimes appear reversed or confusing on the screen.
Is there a way to force math expressions to always display left to right?

[real vessel]

It started happening more 😭

[onyx kraken]

I don't know about forcing. But telling the model what you want greatly increases the odds that the model will do that.

You can also explain, "I want to be clear, we're chatting in Arabic, but I want everything with a formula or numbers to use the world's common notation, left to right"

If there's some other term that's more accurate than 'world's common notation' if you know it, call it the proper term.

I'd give feedback, maybe every 5th to 10th time it does it right, comment that it's doing that right, yay - and each time it does it wrong, allow the model to fix the mistake, show it and let it copy the desired format if needed.

That should tremendously reduce the times it does it wrong.

[little hound]

It doesn't matter how popular something is when it's WRONG. Month is the more important and salient data point, so it should come before day. E.g., if communications were cut short after the first datum, "December...somethingth" is more informative (cold, snow, margin of error no greater than 15 days) than "The 5th of...something" (which provides no information, margin of error is the whole year).

[harsh shore]

How do I check when my advanced voice mode time renews? I closed the popup too early.

[neat nest]

I want to add an AI to my support chat.

[nocturne matrix]

Alright, who's idea was it to make copied contents into a pasted.txt ?!

Models are now taking much longer, because they have to use tools to read the pasted.txt that automatically gets created when pasting a lot of characters

And due to pasting big logs, it creates them with the same name and I've noticed the models are getting confused what to read. Now my solution is to make .txt files with different names so the model knows what to open and read. Gosh.

[heavy cedar]

Hi, I was hoping someone could tell me if using my GitHub connector is supposed to work for me when I try to use it in chats within a particular project I created? I only see it under “more” in normal chats that are not tied to a particular project (so they have absolutely no context). If this is the case, what are the best workarounds that people have used to get all the context from the other project needed to be able to use it fully without having to give me a ton of context for every single request? (I am a GitHub Pro user trying to build my own e-commerce platform for the first time and my setup with ChatGPT chats not having access to my repo slows down everything when working between me, codex and the ChatGPT chats). Any workarounds or best practices would really help

[lime sentinel]

Why did they take away Figma?? Mine just stopped working. Cannot get it to make figma charts for me anymore.

[steep solstice]

I've sent several clips before of me singing or speaking, for the AI to give me advice on how to do a specific accent, or on how to make a specific sound. Usually I just send a video clip with a black screen and it has been able to listen to it, but that didn't work today. Oh well - maybe it's better if I send an actual audio file, so I sent an m4a file - still doesn't work. The AI itself said that I should try making a new chat and send it from my PC, and that I should try mp3 - so I did just that. But, it still doesn't work!

What could be causing this? Is it a bug? Or have they changed anything?

[trail cairn]

there is no audio file support for anything else than code interpreter operations

[steep solstice]

But how did it work earlier?!

[trail cairn]

it most likely just hallucinated the answer

[tawny plank]

Hi

i have a problem with chatgpt

As it wont respond to me

It just keeps loading and loading forever

[trail cairn]

try in incognito mode

[tawny plank]

On the app

iphone

@trail cairn

[trail cairn]

then there are less options like trying on a another network or full uninstall/reinstall

[tawny plank]

Already tried installing/uninstalling

[little hound]

Did they just nerf us AGAIN? Down from 30 to like 10 daily gens? I had 28, then looked two minutes later and it said 8. Geez.

[opaque orbit]

Is anyone not seeing their POSTED content appear on Sora 2 profile? I posted multiple things from drafts and they're not appearing on the profile

[drowsy locust]

Why is month more important even you guy say 4th of July hmmm

[warm folio]

Hi everyone, I've just come accross an issue with Sora 2. Trying to delete a canceled video and it doesn't seem to work, the video stays in my library and doesn't show in trash. I'm using Sora 2 on my browser. Just joined the discored for this matter, any solutions? Appreciate the support 🙏

[vital prawn]

Has anyone figured out a better way to sort Projects? I have a lot of them, but there seems to be no way to put them in alphabetical order or drag and drop them.

[soft crane]

Hello everyone ! Everytime I post a video on Sora it shows it posted but when I go back the videos are gone? Any help much appreciated.

[woeful shard]

Me too

[soft crane]

I hope it’s just a glitch and they will come back.

[lost smelt]

Yeah, a video got removed without warning me.

This is ridiculous.

[soft crane]

My videos came back! I hope yours does too

[little hound]

We also say "July 4th" and "Independence Day." But the colloquial name of a holiday is not the same as optimal everyday usage.

[sudden compass]

I want to start a business. Is there anyone who wants to join?

[heady oriole]

Hi

Guys can someone tell how this instagram account/ influencer

@ayesha.khann_official make so realistic reels?

Please can u guys explain which ai she use to generate this quality? And how she creates consistent character

[iron lodge]

hey guys Ive been doing customgpt for only 2months now honestly im kinda lost, can anyone give tips or sort of tips? on how to instruct them then how afterr?

[shadow dirge]

😎

[livid galleon]

I had to turn off my wireguard based vpn because it slowed down openai connection so much. Anyone else had this?

[late atlas]

why not year first then? is easier for sorting and organize..

[late atlas]

its not.. they re just used to it.. for me best format is ISO.. YYYY-MM-DD for most uses, especially in files and digital world.. for "analog" world, if you want short with numbers the dd'mm'yy is more logical.. and for written with whole words, I can see that month-day can be usefull (December 18 2025), but in my language is same.. we say/use (18 December 2025) cause we dont use "of"

I say is more logical, because you have .... 18 ... (what 18?) of December.. (What december)... of 2025... so 18/12/2025 or 18 Dec 2025.
else it would be December... (what December) the 18th? (eh??? what 18th ?)... of 2025? eh??? xD but i guess is all about habbit

[little hound]

Absolutely not. Year is the least salient, because in 80%+ of use cases, you are referring to the current year, the previous year, or the next year, and everyone knows it. If I said our holiday party was on January 3rd, would you really say, "WHAT YEAR??" No. You'd sound like a dork, or at the very least, pedantic. Year is SO less important, in fact, that it's often omitted. MM-DD-YYYY is vastly superior to all other formats from an informational perspective.

[late atlas]

i explained the reasons.. learn to read silly

and dont start me about superiority and such BS

[little hound]

As did I. Read my previous reply to that.

[late atlas]

it depends on habbit and how you also use it in natural language.. In my language to say "January 13" is wrong... i m not gonna try to explain it more

And ISO format is YYYY MM DD.. for a reason...

[onyx kraken]

I bet one day AI is so prevalent, useful and trusted, that each person sets the format they want to see - and the models involved adjust so each person sees their chosen format.

Nobody's wrong, nobody's misbehaving, we all communicate and the models adjust to ensure we each understand the way we expect to be informed.

There'll probably be language universal software, AI-involved too, at some point, and maybe available notes and disclaimers someone could choose to check about what might be lost in translation/alternate likely meanings.

All with each person just saying meaningful stuff the way they say it, whatever it is.

[ancient ocean]

L sora "We encountered an error while signing you in, please reach out if this issue persists."

[daring tangle]

L

[native stream]

same...

[hazy kindle]

hey guys, my live video with ChatGPT just disappeared today. Any explanation?

i have $20/mo subscription

(please ping if you respond)

[daring tangle]

Are your videos loading?

[nimble talon]

nah its not wrking for me either

[solemn dome]

Please gime Me a óption if the Official app GPT is not avaliable in my country 🇻🇪

[civic stratus]

is there anywhere that i can discuss collaborating with a fellow AI artist?

[charred parcel]

probably but not on this server

[civic stratus]

sad face - seems odd to not be able to discuss working with other artists on a project together

[charred parcel]

there must be Discord servers that allow that, maybe a Google search will help. On this server #server-rules Rule 7 inhibits searching for others on collaboration etc.

[civic stratus]

gotcha - the rule states "to promote non-OpenAI services, products, or projects. " - making it seam like as long as the project pertains to using Open Ai, it would be fine - no?

[charred parcel]

No self-promotion, soliciting, or advertising. soliciting being the key word here

[thick hound]

Phone verification issues. Won't allow me to generate anymore videos when my phone # has already been verified???

[sudden bay]

I think there is a bug with the Thinking model. Example:

A
B
C

would render as:

A
B
C

(on desktop)

On mobile:

ABC

[silk bane]

That is your customized reply style from your past interactions with your AI. I can help you teach it output formatting if you would like.

Encoding and dimensional layers.

Raw Text Example
Single Backtock Block

Block```

```markdown
**This is how you highlight and customize the stdout**

using these as event channels I create a network of links inside the document itself. custom HyperLinks aka SymLinks

[uncut mason]

Hello I have a problem in purchasing chatgpt plus
In the final stage it declined my credit card. In the past I made a purchase for chatgpt plus with this card. I tried several thing unable VPN cookies, etc, I contacted the bank also and there is not a problem. Although it declines the card

[trail cairn]

many have had the same and most have eventually just subscribed via mobile app stores with no other solution

[uncut mason]

Is it a problem that would be fixed? or we do not know what exactly is the solution?

[trail cairn]

there is no fix. afaik most of the time support will say that the user needs to accept online payments or contact bank while bank and setting all say everything is allowed, and yes even with a card that worked before

you can try contacting them or just buy via mobile with the same price

[uncut mason]

so if I pass for example a card to apple wallet and make the payment from them will be accepted?

[trail cairn]

yes

via mobile you can use paypal, prepaid cards, gift cards etc.

[uncut mason]

from the mobile you mean only from the chat gpt app correct? because from mobile browsing i cannot again

[charred parcel]

you can subscribe from the mobile app

[uncut mason]

it say only chatgpt go

it has not for the plus

[uncut mason]

I purchase it from the app
But first the go version and after the plus I could not immediately purchase the plus version from the chat gpt app

[slate flicker]

Is there a way to archive folders?

Sorry I mean projects

[kindred grail]

Does anyone know how to stop ChatGPT from removing lines of code when it updates it?

[analog herald]

you cannot archive the folder aka project . but only the threads in it. When then you restore from archive if the project stil exists the threads restore to the original path. Else they restore to general

[onyx kraken]

Maybe report in #1070006915414900886 , and can show screenshots too if you wish

[wanton heron]

WHEN WILL THEY ADD SORA 2 IN GREECE

[charred parcel]

nobody knows, keep an eye on #announcements for updates

[wanton heron]

aahh okay

[onyx kraken]

And we can watch https://help.openai.com/en/articles/12461230-sora-app-and-sora-2-supported-countries - they just updated it yesterday!

There had only been 7 regions with Sora 2 access/Sora app access but now there's 17!

Not all changes make #announcements , but we can watch the helpfiles too

[fleet veldt]

my acoont got deactived when im enter sora pls help

[charred parcel]

were you using a VPN?

[fleet veldt]

yes

[charred parcel]

from which country?

[fleet veldt]

from USA

[charred parcel]

no, which country are you in now, not the VPN country

[fleet veldt]

ok brazil

im use VPN to enter but this error appearce

[charred parcel]

read this carefully -> https://help.openai.com/en/articles/12461230-sora-app-and-sora-2-supported-countries

[fleet veldt]

I was deleting the data for Sora 2 to create a new one, and when I tried to log in to create an account, it said it was disabled.

[charred parcel]

have you received an email from OpenAI yet?

[fleet veldt]

no

[charred parcel]

well as you read in the link I sent you, using Sora outside a supported country can lead to your account being deactivated. There is nothing you can do about this.

[fleet veldt]

im see now thanks

what a shame

[cedar meadow]

Guys can you give Sora 2 invitation code

[charred parcel]

stop spamming, go to #sora-2-codes

[gloomy jackal]

were there any mentions of when sora 2 will be available in the EU (more specifically Austria maybe?)

[onyx minnow]

this is insane

my chat disappeared

spent 5 hours on it

its gone i have no clue what happened 😭

[onyx kraken]

I'd refresh the page, try searching for some words from it.

If needed, restart the device.

[onyx minnow]

already done i even did an export of all the data

[grand pulsar]

did you maybe archive it?

[charred parcel]

what does this have to do with OpenAI?

[jaunty burrow]

Open AI is powering the application turning a JSON script data into understandable and beautiful result.

[charred parcel]

it's not an OpenAI issue, you can't share links to personal projects here, so please delete them so I don't have to bother a mod

[onyx kraken]

Hey, I'm a community member, like yourself, and I don't buy stuff with credits, but I see this helpfile, https://help.openai.com/en/articles/12642688-using-credits-for-flexible-usage-in-chatgpt-freegopluspro-sora

It has some info about video gens, is that what you're talking about, or some other kind of bought gens?

When you talk about buying 25 or 50 'gens' - are you actually buying 25 or 50 credits?

I see that it's 10 credits for 1 10 second Sora 2 video gen, so 50 credits would get, I guess, 5 video gens.

The usage screenshot, what's that from? I'm peeking at my own Sora 1 and 2, not seeing it. Is it API?

Okay, found it in my Sora 2 'settings'.

It is confusing, I see I can buy a number of video gens for a price, and it doesn't discuss credits.

The helpfile discusses both video gens and credits, and it appears to use video gens as a unit.

like, the 'Sora 2 Pro, High resolution, 15s' - costs 50 video gens.

So, if we made 1 of those, we'd 'pay' 50 video gens to make just one of those.

But if we made 1 'Sora 2, 10s' video, those cost 1 video gen each.

So, when you bought it at the start, what appeared to be however many gens for X price - you started off initially with fewer than it said you'd get - or you use them faster than expected, or not sure?

[broken pebble]

whats happened to sora / chatgpt image generation recently?
it felt like it all became so disgusting and terrible over the last week

top is the old model
bottom is the same prompt on this new model
look how grainy and over detailed and disgusting the colours are in comparison it feels like this is running my local stable diffusion model from 3 years ago

[onyx kraken]

Interesting. I don't do much with Sora gens, so I don't personally know much about this.

I would track how many gens I actually get.

You mentioned you reached out to OpenAI tech support, and they say you have the expected number of gens - if that's the case then the issue is that you don't see the real gens available to you (worth a bug report, feel free to show it, with screenshots if you're willing, in #1070006915414900886 - others may be affected too).

The problem may or may not exist as you use them, keeping track may be a way to tell, and I'd reach back out with the evidence of whatever happens, to tech support again, if it's not the correct limits

[polar matrix]

is there any seo or marketing ai or ideas here anyone can share like how to use sora and other updates

[fathom night]

Hi, I'm kind of new at this but I have a crazy question to ask.
Assuming that i spend a fair bit amount of money just to build a rig to build enterprise systems. Sass whatnot or enterprise systems. Assuming that I can only vibe code. Is it actually possible to do it?

[charred parcel]

ask this in #ai-discussions as this channel is for OpenAI specific help

[random flare]

Hello, I've been having a problem since yesterday when I try to subscribe to GPTPlus by clicking on “Get Plus” on the subscription page. I get an error message saying, “An error occurred on the payment page. Please try again. If the problem persists, visit help.openai.com.” I have tried this on several browsers, without a VPN or ad blocker.

[charred parcel]

do you have an iPhone?

[random flare]

Yes this problem occured when I try on my Iphone and my PC

[charred parcel]

install the ChatGPT app on your iPhone and subscribe inside that using In App purchase

[random flare]

I tried it but on the Iphone app I can only subscribe to Go GPT

[charred parcel]

oh really?

what country are you in?

[random flare]

France

[charred parcel]

are you sure you can't subscribe to Plus in the app?

[green summit]

If the app doesn't show, then go to the browser > chatgpt. com > sign in and click your pfp > upgrade > subscribe.

But it should show in both the app and web usually

[random flare]

On my PC or on my Iphone ?

[green summit]

Any browser, whatever is easiest for you.

If for some reason on Safari it doesn't show properly, click Aa in the top left to enable "Desktop site", and then try again.

Or whatever it's called on IOS

[random flare]

Sorry i'm not english what does that means "pfp" ?

[green summit]

Profile picture

[random flare]

Ok thanks

[random flare]

Thanks but the problem persists I always have this error message "The payments page encountered an error. Please try again. If the problem continues, please visit help.openai.com."

[green summit]

Hm, that's odd. You might have to send an email to OpenAI to get it sorted out. Tho, no guarantee if you'll get a response in time. Seems to be a pattern.

[random flare]

Okay, I'll do that then. Thanks anyway.

[green summit]

No worries

[quaint smelt]

Would you stay? Paying for "Plus" but got 0 service for 30 days. I’m at a crossroads and want to know what you guys would do. Here’s the situation:
1 Month of Failure: Prompts stay "stuck on sending." It’s broken on Web, Desktop app, and Mobile.
Broken UI: I can’t even open Account Settings without it freezing.
Ghosted: Tried Discord, zero replies.
The Irony: I finally went to cancel today and the system offered me "One Month Free" to stay (see image).

The Question: Would you keep a "free" subscription if the service hasn't worked for a month? Is it worth waiting for a fix, or is this just a sign to walk away?
Has anyone else been stuck in this "paid but broken" loop lately?

[charred parcel]

ChatGPT is working for me, very well. There must be something on your end causing these issues because apart from a little instability from time to time the service works.

[quaint smelt]

I appreciate the perspective, Robert! I’m glad it’s running smoothly for you. I’ve been trying to narrow this down for a month, and while I’ve tested across Desktop, Web, and Mobile, I did notice one pattern: the service seems to stabilize when I use a US-based VPN.

Since I'm connecting from outside the US, this suggests the issue might be regional routing or a specific CDN node rather than my local setup. Given that the UI (even Account Settings) is still timing out on my local connection, is there a known issue with regional access or a specific way to report these routing failures to the technical team?

[charred parcel]

what country are you in?

[quaint smelt]

Israel

[charred parcel]

what OS are you using on desktop?

[quaint smelt]

I’m running Windows 11 on the desktop.

However, I want to highlight that the exact same behavior, UI timeouts and prompts getting stuck, is happening on my Android mobile app as well. Since it’s occurring across completely different operating systems and hardware, but stabilizes the moment I hop on a US-based VPN, it really feels like a packet loss or routing issue between the local ISPs here in Israel and the OpenAI edge nodes.

[charred parcel]

bear with me, could you please do a fresh install of Firefox on your PC and try with that?

[quaint smelt]

Tested Firefox: 1-2 minute delay for a simple prompt. Brave/Edge: Still completely stuck. US VPN: Instant.
This confirms it's a regional routing/CDN issue from Israel, not a browser bug. A 2-minute lag on a paid account isn't sustainable. Is there a way to escalate this to the network operations team?

[charred parcel]

you will need to contact support at https://help.openai.com - use this guide if needed https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[quaint smelt]

Thanks for the help and for walking through the troubleshooting with me, Robert. I’ll head over to the help center and open a ticket now. I’ll make sure to highlight the VPN/routing findings so they have a good starting point. Appreciate your time!

[charred parcel]

no worries, good luck with it all

[sullen vigil]

Does anyone know what’s happening? Whenever I send a message the chat gets stuck on loading or throws an error.

I also tried with other Services like Gemini and deepseek.they are working fine and my internet is also stable

Here you can see the chat was giving response normally but suddenly it stopped working and gave an error

[outer kayak]

I am not able to post my video in #sora-submissions please help. When I post my video it gets removed instantly why?

[coarse niche]

Hope this is the right channel for this.But basically i've been having some issues trying to use the voice both on mobile and desktop, neither of them seem to want to actually connect. Anybody else having these issues?

I've done just about every kind of troubleshooting.I can think of on install reinstall login log out clearing cache, clear data et cetera

[cedar flame]

??

[edgy burrow]

Is this a known issue? Getting the same sign-in error

[slim surge]

Uhhh

[onyx kraken]

@edgy burrow I don't see other people reporting they can't sign in.

Steps I recommend:

If you can, try on another device, if you can that helps confirm that the issue is device-related, not your account. If you can try on another network too, that's potentially useful in case it is a network-related issue.

If you can't check with another device, or decide it's device-related, try:

restarting the device
remove any browser extensions (just turning off may not fix)
Clear cache

Any or all of those steps have helped some of the folks that have discussed having that type of issue.

If none of that works, you can reach out to OpenAI tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support

If you do have an account issue, you'll need OpenAI's help.

[tight birch]

How to fix this?

[velvet oracle]

Yeaa

I'm only getting this error when using GPT for chat via Wi-Fi: There's a problem with your request. (9b12f141dccde34b-IST)

How to fix?

[tight birch]

Same

[velvet oracle]

@tight birch and only wifi

Right?

[edgy burrow]

So frustrating .. connected to WiFi too when I had the error
Had it working using 5G mobile connection outside

[onyx kraken]

Perhaps see if resetting the router helps?

[velvet oracle]

@edgy burrow

Important note : This is not a guaranteed solution. The issue likely still exists on OpenAI’s side.
Before proceeding, make sure to:

Set your device’s MAC address to “Stable Device MAC”

Disable IPv6

Steps that worked for users experiencing this issue:

1. Reset your Wi-Fi network.

2. Reconnect to your router and verify access via the web.

3. On mobile devices:

Go to Wi-Fi settings

Forget the current Wi-Fi network (or all saved Wi-Fi networks)

Reconnect to the network

After completing these steps, the connection should return to normal.

[sudden compass]

My network is fine, but it keeps showing Network connection lost. Attempting to reconnect...

[sudden compass]

Why is it like this?

[sudden compass]

always lose the connection halfway through thinking

[sweet crystal]

Before ChatGPT 5, the Study and Learn method in ChatGPT had a special way of teaching. It explained the information in a simple way first, then asked one or two quick questions. After that, it discussed the same idea again but using a different style, like a short discussion. Then it returned once more and explained the idea in a more focused and simpler way. This process helped me understand the information clearly and remember it easily.

However, now the situation has changed. My question is: how can I go back to the old Study and Learn system that focuses on step-by-step explanation, discussion, and repetition to help with memorization and understanding?

[reef warren]

I keep getting Cloudflare 403 errors in the Web, CLI and VS Code extension (It happens more often in the second message.) Is anyone else experiencing this problem? It has become much more frequent in the last day or two

[onyx kraken]

I don't know of a way to 'force' that, but I would discuss with the model exactly what I wanted it to do, how to teach, explain that format, even show it previous chats as a template to use

[sudden sparrow]

Huh

[sudden sparrow]

The web version works fine

[green summit]

For App:

Go to the app/homepage on your phone > hold your finger on the GPT app icon > select "App info" > find storage > "clear data" > reopen the app > sign in > try again.

(Don't worry, none of this has an effect on your account data)

[lucid osprey]

Hi, I just got the plus but I haven’t used it for like a year. Just to verify do we have a chart with all the limits for each task like normal messages, research prompts, quizzes, uploads, etc?

[left fern]

Hi how can I fix this issue ? So far what I've tried is
• reinstalling the application
• clearing the cache

Edit : I connected the VPN and it worked

[teal ivy]

gang i really want to join Sora 2 in desktop BUT IM IN INDIA so what do i do 😭

[wintry parcel]

Hi, there. These articles from OAI help center might be useful:

https://help.openai.com/en/articles/6950777-what-is-chatgpt-plus

https://help.openai.com/en/articles/11909943-gpt-52-in-chatgpt

[pastel river]

I bought the $99 lifetime special for my husband as a gift. Is there anyway I can get a gift card certificate for it? Something to print QR code to a video… something as a gift to put under the tree

[trail cairn]

what lifetime special?

[knotty zealot]

Can anyone help me with this issue? I can't understand whether the problem is with Notion or ChatGPT

Why is Notion in ChatGPT apps not even showing when I try to chat with them?

I don't know what's wrong with Notion MCP inside ChatGPT apps (prev. known as connectors). I tried connecting to Notion via ChatGPT apps three days ago. I verified the OAuth pages, and everything got connected with ChatGPT. But when I tried to chat with Notion via ChatGPT. It wasn't even showing up. Every other app, like Spotify, Vercel, lovable are showing and working as expected.

I thought of disconnecting and reconnecting to Notion, but there isn't any option to do so. Is anyone facing this issue? How do I resolve this? I've been trying to solve this issue for the last 3 days, but couldn't find any solution

[wintry parcel]

Hi, there! I just checked the apps list in my ChatGPT account and I saw "Notion" listed under "Productivity" category. So the issue you are experiencing may be something regional or local? I'm not sure.

Also...I've never used Notion but had a similar experience with another app inside ChatGPT (working fine and then suddenly gone). I think that there was some sort of issue with the tool link, but it got fixed a few days later and it's working fine again. If the issue with Notion just started, it might be worthwhile to just wait for a few days. 🤞

[main ingot]

How can I bypass the copy right characters rule

[onyx kraken]

Any bypass of a rule breaks rule zero, #server-rules by breaking OpenAI ToS and usage policy.

However, patience may work well, because deals are being made that will allow some intellectual property to be used with OpenAI services.

Here's an article about it, https://openai.com/index/disney-sora-agreement/

But until that and any other deals begin, waiting is very wise.

OpenAI, per ToS, if we break the rules, OpenAI could choose to not do business with us anymore - and that could be permanent.

But this is what's coming, and the expected approximate start time (I recommend wait until it is announced as being available):

Among the characters fans will be able to use in their creations are Mickey Mouse, Minnie Mouse, Lilo, Stitch, Ariel, Belle, Beast, Cinderella, Baymax, Simba, Mufasa, as well as characters from the worlds of Encanto, Frozen, Inside Out, Moana, Monsters Inc., Toy Story, Up, Zootopia, and many more; plus iconic animated or illustrated versions of Marvel and Lucasfilm characters like Black Panther, Captain America, Deadpool, Groot, Iron Man, Loki, Thor, Thanos, Darth Vader, Han Solo, Luke Skywalker, Leia, the Mandalorian, Stormtroopers, Yoda and more.```

[stray scroll]

Guys I have an issue. All the account that I make, no matter what and weather I use a VPN or not, All the models Output Using the 4o Model.

[onyx kraken]

I wonder if a browser extension could be an issue, and some of those need to be removed, not just turned off, to check.

Otherwise, OpenAI tech support, can reach out to them https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[stray scroll]

Alright thank you. I will try. Could it be an adblock?

Its so strange btw. Doesnt matter what model I use. 5.1, 5.2 Pro, 5.1 thinking, it always redirects to 4o mini.

[onyx kraken]

I don't know; if the adblock is a browser extension then potentially yes. Over the years a lot of issues with ChatGPT, including weird stuff, people have said they found removing some or all of their browser extensions fixed, and reinstall the extention, problem back again.

Even before there were multiple models we could pick between, it's one of the 'surprisingly many problems seem to come from this, might want to check'.

[stray scroll]

Alright. I will try in some other browser where there is no extensions and i will see if that fixes it

[onyx kraken]

Another thing worth trying, potentially, is a fresh install of Firefox browser. If you want to check, before you dive into OpenAI tech support, that too seemed to sometimes fix odd problems, though been a year or more since I last saw that discussed

And if you can check on another device, that might help indicate if it's something device specific, account specific, help narrow it for tech support if you need them

[stray scroll]

Alright Thank you

alright the browser with no extensions work. I will try to figure out which extension it is that is breaking it now

Okay no. @onyx kraken Could it be browser based? The edge browser works perfectly fine however the chrome even with all extensions removed doesnt work at all/

[onyx kraken]

Very interesting.

I use Chrome as my browser and I don't seem to have any issues with auto being forced to be any particular model.

So, I don't think that Chrome is straight out broken for everyone.

However, you could have cache involved, or any number of other specific to your system issues.

I'm not a computer expert, I just remember what I've seen people discuss (that I understood) that helped fix stuff, so I can't really explain or reason through this, just point out what I noticed others have said.

[stray scroll]

Oh what

Should I clear my cache and cookies for chatgpt

and then log back in

[onyx kraken]

It might matter, and might be more than chatGPT, like if a browser extension has set stuff for the browser itself? That may be an issue. Again, I theorize, I'm no expert here, just remember what others have said

[stray scroll]

This is what happens for me

[stray scroll]

holy thank you so so much.

Clearing the data was the issue that was required

You deserve a raise tyyy

[onyx kraken]

Hehe, I'm a volunteer community member, like yourself!

Really glad you were able to help fix your issue.

[stray scroll]

You should be getting paid. You help us way better than the support ever well tbh

[onyx kraken]

Thanks. I'm sure OpenAI tech support can help with stuff I couldn't even begin to solve - but for folks who need help with something they could personally fix, just need to know what to do? That's one of the reasons why we're here and chatting with each other 🙂

[civic stratus]

when trying to open a conversation: i get a message 'Unable to load conversation 69451be9-52a8-8332-98f3-d84998b3d5f5' - has this happened to anyone? i have tried refreshing, clearing cache and disabling extensions

[onyx kraken]

Sounds like a chat that became buggy. That's more common with really long ones, and I'm not aware of any fix.

I tend to keep my chats on the shorter side, and make a new one as soon as I notice lag or anything else that makes me think it might start to bug out.

[civic stratus]

interesting - okay thank you

[civic stratus]

do you know other way that i can access the thread to review the conversation? stored elsewhere

[onyx kraken]

You can try requesting your data export, https://help.openai.com/en/articles/7260999-how-do-i-export-my-chatgpt-history-and-data

I am not sure that can rescue a bugged out, not openable chat, but it's a great thing to do every so often anyway, just to help back your data up.

[civic stratus]

i appreciate you

[onyx kraken]

Appreciate you too! I hope you can save your chat

[tough halo]

Hi I wanna use sora freely so what I should do?

[misty falcon]

Hi, can someone here help me regarding something related to reddit?

[peak granite]

can't enter into my old Sora page. it asks my birthday and that's it...no button to logon...please help, i need access to my page asap. thank you

[onyx kraken]

That's happening for you on https://sora.chatgpt.com/explore ?

[peak granite]

this appears and i can't login anywhere

[onyx kraken]

Can you remove the onboarding part of the url, and go to sora.chatgpt.com itself?

[peak granite]

i did it and it's the same

[onyx kraken]

You either need the help of OpenAI tech support, which you can reach out to through https://help.openai.com/en/articles/6614161-how-can-i-contact-support

Or it may be a device specific issue, like a cache/browser extension-related weirdness.

If you choose to explore fixes yourself, I would try to connect first from some other device. A phone or tablet (you can use the browser to connect, doesn't have to be the app, which may or may not be available where you are anyway - but I think you're showing browser, so use that) or if you have access to a different computer.

If you don't get this weird screen from another device, then your account itself is fine - the issue is something about that device.

To handle that, I would remove (not just turn off) browser extensions, if there are any (Surprisingly many weird problems have been avoided by that, based on what people discuss about problems they've had with ChatGPT).

I would clear cache, and restart the device.

If that didn't fix it, I would try a fresh install of firefox browser (been a while, but that was reported to be a fix for a range of problems at one time, never knew why but that seemed to work when some other fixes didn't).

Good luck, sorry you're affected.

[keen portal]

Good afternoon, I have a question. When ChatGPT gives me an answer to a question and that answer contains <svg> code that should display an image, instead of the image itself, I see a broken image icon.

I tried two browsers, Google Chrome and Opera, and the problem persists.
I wanted to clarify, does this happen to everyone? If not, is there a solution? Thank you.

[hollow wharf]

I hope I can explain this correctly. I am writing a story. I have it in word and I'd like to have ChatGPT read it over. It is a big file. I'm not sure how many words it is but it's in the K10s. Do I have to rewrite it into Chat GPT? Will it be able to handle that many words? And is there a way I can just upload a .doc or .docx file to ChatGPT somehow?

[hybrid zinc]

Where do I go to um report an issue with my year in review 🤔 bc I got something F****D in mine... and I really think they need to be more careful about what they choose to highlight...

[charred parcel]

https://help.openai.com

[compact drum]

Guys I can't decide my major if it should be software or AI or cyber security

[upbeat plover]

This is what i'd do.

Select all, copy, notepad, paste, save, upload .txt to chat gpt. It can take a variety of documents if you have a plus plan. So maybe you can just upload the .doc

[hollow wharf]

Thank you so much. Worked like a charm. Cheers.

[echo helm]

Hello all. I wanna ask do anyone know how to shorten the seat of a GPT-business space

[stray scroll]

Hello. Whenever I attempt to enable the "Use additional Models" on the Buissness plan it always gives the error "Unable to update Feature setting". I checked the console and it told me "401 (Unauthorized)" to the api url it had there. What?

[echo helm]

btw I kept failing to login the openai website, have anyone been in this situation before?

[stray scroll]

wdym by failing? whats the error?

[echo helm]

I login and the website still ask me to login

[stray scroll]

uhh thats strange

Yeah sorry never seen that one before

Try clearing all your cookies for that site

[echo helm]

i tried this earier but it didnt work

[old wigeon]

Hi everyone! Has anyone encountered the issue where the thinking mode doesn't work? I make a request, and it responds instantly, without even showing that it's thinking. However, if I repeat the request on my phone, everything works fine

[sudden compass]

Why is there no option to open the camera in the live chat anymore?

[wintry parcel]

Hi, there. Have you tried a different device? Change browser?

[thorny hollow]

Anyone have solution for this?

[green summit]

#sora-discussions message

[woven bobcat]

Anyone know what this means?

All I tried to do was open the app

[swift sorrel]

yes this is maybe because you change ur password or smth

happened to me and i just had to re log in

@thorny hollow

[thorny hollow]

But how i didnt changed anything🤔

[woven bobcat]

Ok apparently its not my fault 🤷‍♀️ guess ill try all the suggestions again and just wait 😒🤣

[thorny hollow]

@swift sorrel

[woven bobcat]

Nope nothing worked 🤣 ahhhh I use my chatgbt for so much I feel lost right now 😭

I made it worse 🤣

[desert cloud]

app works but cannot login on computer via browser, no matter how hard I hit the login button does nothing, but openAI status says all good??

happening on all browser's even via VPN so its not my net, something up with their CDN?

[desert cloud]

it could of been cloudflare it wasn't refreshing my token properly..
*but still wont load any of my chats 🙁 *

[icy gyro]

Hello is anybody else having those problem with osx and Chatgpt Chat Bar

basically after last update i have two problems sometimes when i launch it i have this where i cannot type properly the placeholder text is still there

and the other problem is that after opening it the focus is not set to the text so i cannot type right away i have to click with mouse into it

i tried reinstall, with cleaning all the caches

does not help

This was for me a killer feature for chatgpt, and without it its a huge bummer.

[swift sorrel]

that happens its common jsut re log in

[cerulean crypt]

Anyone else having issues with chatGPT? My chat history is gone and unable to prompt the engine

[arctic cedar]

nope

[icy gyro]

what are you running

im on 26.2 (25C56)

[arctic cedar]

26.1

[icy gyro]

i tried everything that i can, clean install did not help

[arctic cedar]

what app of chatgpt r u using beacuse im just using the app

[icy gyro]

Version 1.2025.343 (1765590282)

[arctic cedar]

what version am i running

i think im more advanced

[hollow crow]

what can i do ?

[onyx kraken]

Here's some info about it, https://help.openai.com/en/articles/9101499-what-happens-if-my-business-workspace-is-deactivated

https://help.openai.com/en/articles/8798618-what-happens-if-a-workspace-has-been-deactivated

https://help.openai.com/en/articles/8266418-data-retention-when-a-member-is-removed-from-a-workspace

https://help.openai.com/en/articles/9295112-memory-faq-business-version

https://help.openai.com/en/articles/8801890-can-i-migrate-or-merge-my-personal-chatgpt-workspace-over-to-my-chatgpt-business-workspace

https://help.openai.com/en/articles/9123978-how-can-i-leave-a-deactivated-business-workspace

If you're the owner, check payment and reach out to OpenAI tech support if you need help, https://help.openai.com/en/articles/6614161-how-can-i-contact-support

If someone else owns that Business account, and has chosen to remove you from it or has not continued to pay, your options are pretty limited to regain access, but I'd reach out to that owner and discuss

[hollow crow]

yes but the problem is that i cant choose a new workspace or account

[onyx kraken]

Does this help? https://help.openai.com/en/articles/8798594-what-is-a-workspace-how-do-i-access-my-chatgpt-business-workspace

If not, what point is not working?

[hollow crow]

well as you see there #community-help message, theres no workspace a can select or create, and i cant disconect from my account. no button i can click or anything.

[onyx kraken]

If you can't click anything, you will need to reach out to OpenAI tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support We're community members here, like yourself - none of us can adjust your account from the inside if something's bugged.

I notice the URL you show has more than https://chatgpt.com/

Might want to try checking the url directly, and see if there's any way you can open your profile info. For me that's bottom left (but I'm an individual account), helpfile says top right so check either maybe.

[ancient shard]

does anybody else have problems with billing when u have sufficient funds?

[echo helm]

I've tried google chrome & safari both same

[wintry parcel]

Disable VPN (if using)? Try different device/ network? These are typical troubleshooting steps (along with cache cleanup and different browsers, which you already tried).

[patent bear]

So i just subscribe to ChatGPT Plus and my Sora 2 video limit are just 10 instead of 30? what's going on? can someone explain what is happening?

[charred parcel]

how do you know your Sora 2 limit is 10?

[patent bear]

because it says in my usage page that it has only 10 video limit per day.

[charred parcel]

ok, well the best thing I can recommend is to contact support at https://help.openai.com

[patent bear]

Okay. It came back to 30 videos per day thank god. So i just had to wait for a couple of mins for it to update.

[stray scroll]

Hey guys I swear I this is annoying. No matter how many accounts I make they are always banned after around two days. It happens when I pay for one of the subscriptions. It doesnt matter what subscription it is (plus or teams). Doesnt matter If I don't even use the account at all. Its deactivated with no reason output in the ban email. Im so confused 😭

[charred parcel]

if you are banned on one account it is against the ToS to make another account

[stray scroll]

Right, is there any way to appeal this? because im wanting to subscribe to chatgpt plans.

[charred parcel]

https://help.openai.com is the way to get support

[stray scroll]

Alright

Thank you

[charred parcel]

this is not a place to look for work

[cobalt kestrel]

Can anyone direct me to a way for agents to consume SDK/API documentation for Agents? at the moment I am using a very crude method, ligthpanda. wanted to know if there is something else. MCP, or maybe creating local cache with semantic search tool?

[cobalt kestrel]

Don't be upaset because AI told you there must be a god.

[past tapir]

All my previous chats from the side bar are gone. Any idea what happened? How can they be restored?

[wide terrace]

Hi,In which chat do I type prompts to create images? thanks

[charred parcel]

just prompt like create an image of a cute kitten

[ancient herald]

why was my message deleted?

it's not about Sora FFS

I had a link for inviting peopel to use chatgpt plus for 3 months for free.. but I can't find it anymore

i saw it in the web website.. and it said it would have been up through jan 2026... now i cant find it any longer.. where can i get it again

[onyx bough]

does anyone know if pprojectt is for free or paid version? reply to this message ty

[onyx kraken]

You can discuss any moderation action with modmail, #safety-n-help shows how

[onyx kraken]

"Projects are available to all free and paid subscription types globally." https://help.openai.com/en/articles/10169521-projects-in-chatgpt

[civic stratus]

Anyone know how to get this solved? Been happening for 2 days

[charred parcel]

are you using a VPN?

[civic stratus]

I am not

[charred parcel]

if the message keeps occurring contact support at https://help.openai.com

[empty ivy]

Is everyone making Sequin outfits slowing CHatGPT down to a crawl? I'm on Pro but I have to wait about 5 minutes for a simple response now...crazy

[onyx bough]

anyoneh ere to help me?

[onyx bough]

what happening?

this is project???

in*

i am oly getting this there

is it quiet here all the time or just now?

[gilded dune]

Team where do I find support for this

[onyx bough]

most questions is pretty much help.ChatGPT.com

[onyx kraken]

GPTs, CustomGPTs are not the same as projects.

I don't know if you're using one in a project, but projects themselves, with whatever limits, are available for all members, per the helpfile I linked

[gilded dune]

This is such a pain tbh

Can't pay to ujpgrade

lol

[onyx bough]

can someone confirm if im not isong GPTs or custom GPTs

[gilded dune]

Also out of curiousity WHY does GPT not work on my chrome, like ever

[onyx kraken]

Hey hi, we're community members here, like yourself! The very few of us who work for OpenAI have gold names, #server-staff. But even those folks can't help you in this forum, you'd need to go through tech support who have the ability to access your account if needed. https://help.openai.com/en/articles/6614161-how-can-i-contact-support

Sorry you're having trouble, hope they can help you soon!

[onyx kraken]

Well, you linked this screenshot?

And that says you reached your limit for using 'GPTs', which as far as I know means CustomGPTs.

[gilded dune]

Did, any idea why it doesn't work on chrome?

[onyx kraken]

I'd check to see if I had any browser extensions, if so, remove - not just turn off, people have reported that removing is needed for a problem one, not just off. If using VPN, turn off.

Reset cache, clear cookies, see if that works with Chrome is my best advice, good luck!

[onyx bough]

i got this You are correct—earlier versions of ChatGPT and GPTs allowed Free users more access, but starting mid-2025, daily usage limits were set for Free users due to the expansion of advanced models and features. Now, Free users can only use GPTs a limited number of times within each five-hour window. This change was made to manage demand and ensure fair access as capabilities expanded.

i am wondering if its hallucinating

[onyx kraken]

I use Chrome for ChatGPT, it's my preference, and I don't personally have an issue, but that doesn't help narrow down what your issue might be. I would try those areas though, those are the normal problems when someone says they found a fix

[onyx kraken]

You can ask in #1184149552639516732 , which uses a model trained to use OpenAI documentation to answer questions, but -

This helpfile https://help.openai.com/en/articles/8554407-gpts-faq says,

"Free tier users can discover and use GPTs in our GPT store with message limits."

"Is performance for GPTs different for users on the Free tier?
GPT usage on the Free tier is subject to the same limitations as ChatGPT. GPT advanced functionality includes data analysis, file uploads and web browsing. As a Free user, you may hit tighter limits for advanced capabilities. When you hit these limits, you may see a message such as the one below:"

[gilded dune]

No live-chat support either and it's chirstmas so I'm not getting a response for w hile

nice

[onyx bough]

i ddon't think there ver was live support

[onyx bough]

im not using GPT store im jsutjust using simple projectect sectio...

[sharp epoch]

is there a prompt which helps chatgpt make better clash royale decks

[plain tartan]

I was banned from chatgpt my account
I appealed it but since then no reply from them
Anyhow or where i can chat a chatgpt employee or call customer care to help me get my account

[onyx kraken]

Usually when an account is suspended, there's an email sent to the linked email account, and that usually has where would reply to appeal, I'd read that email carefully if you have one and follow any steps you are invited to follow.

There may be a delay, this is a major holiday in the country where OpenAI's main offices are.

You can also reach out to OpenAI tech support through https://help.openai.com/en/articles/6614161-how-can-i-contact-support if you would like to do that.

[charred parcel]

were you banned for using Sora 2 out of allowed countries?

[plain tartan]

I noticed this a common issue
Do people really get their account back
And could you help with this i want to request a refund will I contact Google to refund my money or Chatgpt?

[plain tartan]

Nah

[novel birch]

In Agent Builder, it seems like the only option to output a message like "Let me know if you need any addtiional help" after displaying something like a widget card is to spin up an Agent node. Am I missing something obvious that would allow the system to output a static message to the user?

[onyx kraken]

https://help.openai.com/en/articles/7232895-how-do-i-request-a-refund-for-my-chatgpt-subscription

I'm a volunteer community member here, like you! I know only what I personally experience and what people tell us, and they don't always tell us everything - especially about problems getting fixed. After all, once it's fixed, they don't need to ask for more help, and that might have been the only thing they were here talking about.

However, across the years, I've seen probably close to 1-2 thousand folks discuss account suspensions, and I have noticed only 1 who talked about getting access back - there sure could be others; not everyone discusses everything and I don't read even half the posts folks make here.

I recommend everyone know and follow both https://openai.com/policies/row-terms-of-use/ and https://openai.com/policies/usage-policies/

OpenAI discusses how it handles enforcement in those, and in https://openai.com/transparency-and-content-moderation/

They also offer https://openai.com/form/usage-policy-update/ for people who would like to be directly informed when usage policies change.

But as we look at how OpenAI enforces:

I've seen a lot of people discuss warnings too. I recommend to everyone, know the rules, and consider following them. Breaking them, once noticed, could cause OpenAI to choose not to do business with any of us in the future, and that could be a long time to regret the choice to break rules, even ones we didn't like or never wanted to follow.

They also state this about the appeals process,

[sour tusk]

Do those rules apply to all account tiers? If a paid account is suspended, can the user cancel the subscription?

[charred parcel]

the rules apply to all account tiers and if your account is suspended the subscriptions tied to the account are cancelled

[plain tartan]

What is wrong with chatgpt
I paid for a subscription tier
They haven't gave me and still asking me to pay again
Worst platform ever

[fluid kelp]

heyo

[empty ivy]

Can anyone else confirm how slow chatgpt is rn? I can't even use it for my work, how much do I have to pay to get good service?

[vivid mortar]

It looks like ChatGPT is fully operational right now, what are you seeing? https://status.openai.com/

[fluid kelp]

will sora be monetized or no?

if it will then when/

[rugged wing]

Where do I go to thank someone from open ai not a Ai response?

[onyx kraken]

This discord is mostly community members interacting together. The few here who work for OpenAI have gold names, #server-staff , but it's pretty rare to see them post and it's not clear that OpenAI tracks our flow of chatter to each other here.

Two channels that are not ideal for a thank you, #1070006915414900886 and #1070006151938314300 , do seem intended to be easy for OpenAI to look over whenever they choose (and sometimes we see gold-named folks replying to ask for more info or otherwise discuss in those channels), but that's not really a place for a thank you.

Maybe #1155775326253756456 if you want to talk about how it specifically helped in some personalized way, but remember not to share PII or other stuff that a random person shouldn't know, because most of us here are community members who choose to be here, and there's no way to ensure there are no folks with ill intent.

[empty ivy]

Thanks yeah I do not get it. it says status is up etc but slower than molasses, supposedly all the computing is done server side so does my PC/Browser have that much affect?

[vivid mortar]

Which model are you using?

[left trench]

Hi im new to the server and I'm working on finding and training an Audio classification model for a project then integrate that model into a website being developed by my partner. The way I wish to do this is by using data online and using it to train the Wav2Vec2 model on hugging face. I'm classifying bird calls for a range of species common in my area. Is there any advice or recommendations you could give to help me.

[empty ivy]

I've had similar issues on 4o but I'm now on 5.2 but also on "auto" think.

[ember citrus]

How do I get Sora 2 if I live in Sweden?

[onyx bough]

What did you do?

Can someone please help me with something? I just want to know if on free tier project has unlimited prompts I know someone sent me an image over here stating that it is free but I don’t know if this is regionally locked because I cannot send multiple prompts on projects which is frustrating. It’s something I want and if I can’t have that I might as well use a different platform.

[onyx kraken]

I'm not personally free tier right now, so I can't confirm -

But this helpfile https://help.openai.com/en/articles/9275245-chatgpt-free-tier-faq states

Free tier users can use GPT-5.2 only a limited number of times within a five hour window. We’ll notify you once you’ve reached the limit and invite you to upgrade to ChatGPT Plus.```

This helpfile https://help.openai.com/en/articles/10169521-projects-in-chatgpt confirms that free members can use projects, including sharing them, this limit for uploads is mentioned,

```File upload limits vary by subscription type. Note that only 10 files can be uploaded at the same time.

Free: 5 files per project```

```Collaboration Limits

Free users: up to 5 files and 5 collaborators```

It does not specifically discuss message/prompt limits, but since free seems to have limited messages in general, likely limited inside projects too

[covert charm]

How do I fix this? The app is fully updated and I already tried logging off and back in

Nvm I fixed it

[nimble steeple]

I wanna know...Is AI truly as bad as everyone makes it out to be?

[onyx bough]

yes top fortune 500 companies are not hirring

[nimble steeple]

Huh? No I meant like...are we destroying the environment...is ChatGPT and Sora supposedly destroying artist, and creativity?

[white ingot]

I'm a Designer and -> no, AI ist a tool in my Workflow. a great one.

[midnight hearth]

The resources dumped into them could certainly be better used.

[nimble steeple]

Thank you. Clear cut answer for me. Ive just needed like confirmation because I've been feeling like the only sane one, and been pretty bummed...And just... Ive recently got harassed by a traditional artist for 3 days and now has gotten to a point I have to bring law enforcement

[wintry parcel]

I'm sorry to hear this. There are also many artists, musicians, filmmakers and other creatives who are excited to explore the creative potential of AI.

[nimble steeple]

I want to believe there a world traditional and ai artist can coexist

[rustic marlin]

people will never stop making art

[nimble steeple]

I would hope not. Sorry just really im my head, this artist has slandered me so bad that she has randoms commenting on my videos, post, pages, dms anything with just hate saying im an art thief, im destroying the planet, and killing creativity...

[rustic marlin]

listening to hate isnt productive

[white ingot]

this...Professional designers who ignore AI will eventually be at a disadvantage.

[wintry parcel]

It seems to be happening already. I know a professional designer who says AI is already integral to his workflow.

[white ingot]

in mine too. at many levels.

[noble phoenix]

i really wish more people had that mind set :/

as someone who uses ...ALOT of ai.. i literally pay for like.. 5+ different AI subscriptions .. heh i personally see a lot of flack

[silk flicker]

Does anyone else have the same bug?
ChatGPT search does not work at all, always no results:
Here i searched for Safari which is clearly visible in my chats

[brittle ocean]

Hi guys, Emoji Creation Challenge is over and no longer available?

[green summit]

Correct. They closed it and are now checking the Top 10 (and possibly top 11 - 15). Will be revealed soon.

[brittle ocean]

And one more question: has anyone paid attention to the creation of images using "Image Gen" 1.5? All images are noisy, whether through the API or through the ChatGPT. They cannot be used in normal production...

[solid gyro]

I need help

[dry flame]

hi, when i use a single chatgpt window for a long project to maintain the continuity, it keeps on crashing or takes too long to respond after some time, is there anything to solve that issue

[wintry parcel]

Hi, there. As I understand it, there is a limit to the context window, and as the chat becomes longer, GPT struggles to manage all the tokens. Experienced users have stated they open new chats often to prevent this. Have you tried using "Projects"? You can keep all the chats related to the same project, upload documentation, etc., for continuity.

[late python]

Can someone explain to me the difference between an agent and an assistant? What can each be used for individually? Or perhaps an agent can be linked to an assistant?

[onyx kraken]

Agent is a more defined term, in how OpenAI uses it, than assistant.

ChatGPT itself in all forms 'is an AI assistant'. https://help.openai.com/en/articles/12677804-what-is-chatgpt-faq https://help.openai.com/en/articles/9260256-chatgpt-capabilities-overview

However, in OpenAI's API, 'assistant' is a specific term, separate from what ChatGPT is, https://help.openai.com/en/articles/8673914-gpts-vs-assistants https://help.openai.com/en/articles/8550641-assistants-api-v2-faq https://help.openai.com/en/articles/9242831-how-does-the-launch-of-assistants-api-v2-affect-my-assistants-launched-with-api-v1

The Assistants API allows you to build AI assistants within your own applications. An Assistant has instructions and can leverage models, tools, and knowledge to respond to user queries. The Assistants API currently supports three types of tools: Code Interpreter, Retrieval, and Function calling. Assistants are designed to be created by developers using the OpenAI API.```

Agent is a ChatGPT tool that allows ChatGPT to 'do more things', including complex tasks, https://help.openai.com/en/articles/11794368-chatgpt-agent-release-notes https://help.openai.com/en/articles/11752874-chatgpt-agent

[crisp flax]

Hey all
Is there a way to change my OpenAI account’s email
I started my account with my work email and now I am transitioning to a new job and want to keep my account

[charred parcel]

no, read this https://help.openai.com/en/articles/4936827-how-to-change-your-email-address

[crisp flax]

Thank you

[desert hedge]

Why do they not allow you to change your email address? This is one of the very few services that I’ve come across in my lifetime that doesn’t allow this

[onyx kraken]

Some options:

Data export your chats.

Directly yourself copy memories, personalization.

Be ready with your new email to start over, with the information you want to give your new account, so you can continue as you were, or as close to that as you can be.

https://help.openai.com/en/articles/7260999-how-do-i-export-my-chatgpt-history-and-data

[onyx kraken]

The helpfiles do not provide a reason, simply, "We do not support updating the email address associated with your ChatGPT/OpenAI account."

[mighty garden]

Hi,
Once exported, how do you see your chats in a "normal" /"readable" way? For me is just a huge sausage with 2years of conversations...
Thanks!

[late python]

Honestly, I need a simpler explanation, maybe with an example, because my imagination can't grasp it.

[green summit]

Basically,

Assistant: you send a task > it replies > then waits for your next message.

Agent: you send a task > it can plan and take actions using tools > running multiple steps until it finishes or it needs your input again.

• Assistant: explaining stuff, rewriting text, making plans, generating ideas, helping you decide, answering questions, tutoring, summarizing, drafting messages.

• Agent: doing multi step tasks like researching across sites, comparing options, filling out structured info, cleaning and transforming files, pulling data into a doc or sheet, running repeated checks, following a checklist workflow end to end.

[onyx kraken]

That is how it comes and I'm not aware of a provided one-step way to separate them into individual chats.

Some ideas, using cntrl F to search for keywords of interest and read the chats that have them.

You could also explore with ChatGPT writing a program for you to use that would separate them into individual chats for you. You could also do that by hand if you wished.

[vapid warren]

hi guys, do you think GPT is getting worse than before? I used GPT thinking mode (Standard and Expand) for several months, usually it think some minutes then give me an answer. But now it can't think, it will directly answer my question in seconds and the answers are very very ridiculous. Does anyone have the same issue?

[mighty garden]

Thank you! I thought is something that i was missing 😊

[floral nimbus]

does chatgptGo have the option of seeing your phone camera/video chat?

[golden crypt]

Hey, I requested my data from the OpenAI privacy centre around 48 hours ago but still don't have an email with the export download link? I received the acknowledgement confirmation email and I've checked my spam folder

Anyone else having this issue?

[onyx kraken]

I would ask again. If you don't have it a few hours later, ask again. And again if needed.

For me when I've asked for mine, sometimes I get it 2 hours later. If I haven't gotten it within 4 hours, I ask again. Sometimes I ask again for over a week before I finally get it (system is partially down, maybe?) but I've never gotten one more than about 4 hours after I asked - so if its been that long I don't think it will come from that request - ask again.

[daring quail]

Heeey I finally found a fix for Chat GPT voice weird noises and sudden stops on Android.

Option path: Developer Options > Bluetooth Audio LHDC Codec: Low Latency > Turn ON

[proven swift]

i am having to restart my ChatGPT after every prompt, it seems to go into not responding every time and hangs for up to 15 minutes, tried opening it in incogneto, same thing. I checked connection and server even against other ai programs (night cafe) and they respond fine. any ideas?

[trail cairn]

with a new chat or some old mega long chat?

[golden crypt]

Thanks. I've resubmitted the request 3 or 4 times now, all at various times of the day. It's just stuck on pending/in progress. I have another request with an alternate account 48 hours ago (I didn't cancel and retry for this one just to see if it was delayed).

I emailed support and they just advised me to request it through the Privacy portal (which is exactly what I did and I mentioned this, so not sure why they answered with that), so not helpful unfortunately. I genuinely think it's a problem on their side (there are Reddit threads talking about it too). I've emailed the CEO but not hopeful for a response.

[onyx kraken]

I would still keep trying with new requests, until one succeeds, unless you do not want to. That's what I do myself and what I've heard others do - eventually successfully if it doesn't work on the first try.

[golden crypt]

Thanks, appreciate the reply and yeah will do.

[worthy marlin]

anyone know an app that saves and repurposes the context across llms?

[onyx bough]

is ChatGPT slow for anyone?

[unborn veldt]

For the person who added the ❔
I have images added to a project. (Plus member, memory sharing on) I ask 5.2 what files exists in the project context and it seems to be about 50/50 if it can find the files or read them if it does find them.

[onyx bough]

does premium increase speed? what is the cause of slow speed/

are you on paid?

[serene sun]

Hey guys

Not including that stupid clippy

Everyone is breaking the rules of Sora

[onyx bough]

what do you mean?

[restive copper]

My chatGPT lilbit slow if i continue the talk on history

[oak rune]

wassaaaap i pay for pro and my chatgpt just keeps on reasoning and giving up after making me wait for 15 min does anybody know how to fix it?

[misty kernel]

i'm not sure what the reason for that would be LOL!

[golden crypt]

@onyx kraken I keep getting this server connection error on the OpenAI privacy portal. Could this be the reason why the request isn't going through? I've tried different devices, browsers and networks (WiFi and mobile data), as well as Incognito mode and clearing cache and cookies.

[serene sun]

Ok let me tell you

[serene sun]

People are bypassing with hexadecimal

We’re all gonna die

[onyx kraken]

I don't know, and I'm a community member like yourself, I don't work for OpenAI;

but I've never seen that message. The problem is a common one for folks with issues using OpenAI services, to the point that if someone can't use their account or a feature of it and doesn't know why, "Remove, don't just turn off, browser extensions because even turned off they appear to interfere for some peoples' systems with various ChatGPT and account related things." absolutely is advice I give, and I've seen... probably a few hundred members over the years report that fix worked for them (over a hundred said just restarting their device helped, other standard fixes are clear cache, try on another device, try on another network, and/or try with a fresh Firefox install - that pattern of fixes does fix most account issues that don't need OpenAI tech support).

For me, I don't think any of those fixes made my data export request succeed - and you mentioned getting an email confirming your data export request was made? I don't think I've ever gotten an email like you describe (but maybe that's a new step they added?) But you said, https://discordapp.com/channels/974519864045756446/974519864045756454/1455369931972415509 "I received the acknowledgement confirmation email" - I've never seen that. This is what the only response used to be:

[rugged wing]

happy new year 🥂🎉

[blissful ruin]

Hey, am I the only one that still doesnt have the new UI for voice mode on IOS? There is nothing when I click on the voice settings either, would love some help

[blissful ruin]

Happy new year!

[golden crypt]

Happy New Year! Just over an hour to go here in the UK 🙂

[golden crypt]

I know you're not an OpenAI employee, which is funny as you've been more helpful than their support 😂 I appreciate the detailed thorough replies

I see the error just under the 'your request is being processed' status message near the cancel option. I don't have any browser extensions, ad blockers or VPNs on my device. I'll try with Firefox as you recommended.

Support responded just to confirm that they've received the request and to let them know if I don't get the email within 7 days. They ignored my other questions about why it's taking longer than usual, but oh well. I'll post back with an update either way.

[onyx kraken]

Hope you have it soon.

I haven't exported data in a while, so the process may have changed since I last did.

I would still be tempted to try regularly if I hadn't gotten my export within a few hours of the request, but that's based on what worked before, a conglomerate of what people have said and my own experiences (which are not really current).

[opaque tiger]

I added API credits and used an API token to make a bunch of calls to GPT-5 Mini, I know it consumed tokens, I know I'm on the correct account, but on my project page it says my total spend is $0.00 and that I've made 0 requests on anything at all. The interface is super arcane and opaque and unhelpful. The Logs tab clearly show I made the calls... but for some reason I'm either not being billed for them, or there's a possibility that this usage doesn't even break 1 penny of usage.

But I doubt that because I surely must have made enough calls to break $0.01 of usage - and also those requests should still register on the Usage page even if the total cost of it all hasn't broken $0.01 yet.

The part that's extremely annoying is in a computational interface where everything is supposedly automated, if this is one of those "it takes us 5 business days for our carrier pigeons to verify the requests you did in fact make" this already has intimations of being some corporate bureaucratic experience where you can't build anything on OpenAI's platform without going on some mythical quest fit for an ancient Greek Epic or something.

ChatGPT says GPT-5 Mini is not metered... despite the pricing page saying it clearly costs money to use.. oy vey

Even ChatGPT agrees this is bad design.. and yet I have no idea if what it's saying is authoritative, or if maybe those GPT-5 Mini calls chewed through a billion tokens and the SWAT team is going to bust down my door and arrest me for being a bad boy who went over my credit limit, or what.

[opaque tiger]

Yes... I should probably calculate it and see how much actual cost I incurred, but the Usage page should at least log all the calls I made and show the fractions so that it's clear that I am in fact using my credits

Ah, you can't read. My mistake.

Well now I have two - first this opaque interface, and now I also have you pretending to help; I'm sure more people will pile on. Thanks!

[late python]

What do you recommend? All chats – each topic separately? Or maybe everything in one chat/conversation? Mine is freezing up terribly. I've disabled the thinking function in it and I'm still waiting a very long time for a response. I get the message: "The page isn't responding." I think he got offended and left.

[charred parcel]

long chats don't work well in the browser, so make your chats short or use the ChatGPT macOS Desktop App

[stable jackal]

hi how do I use ts

[hollow badge]

🥀

[stable jackal]

Hey there, stranger!

[dim ermine]

Hi there, what do you want to use? Can you please elaborate

[stable jackal]

??

[stable jackal]

I don't understand how I am supposed to open the ChatGPT application on my mobile device.

[stable jackal]

kk

[worldly quiver]

Hi

[little wren]

Hi - how do I strat the basic of earning

[charred parcel]

what do you mean?

[little wren]

well - basic of getting into project to start earning

[charred parcel]

start earning what?

[little wren]

start freelancing and earning money - Am I in the wrong place asking a dumb question

[charred parcel]

yes you are in the wrong server

[little wren]

makes sense 🙂

[mighty garden]

Look on jobs websites?

[true mauve]

Hey so, genuine question, how long does it take for a payment to hit? I had to update my card to another one and I've tried. It says next billing is the second, well it's 5:21 am on the second, when does it hit today?

Hell I've even tried rebuying plus

[fervent gull]

salut chat

[rugged wing]

@chatgpt hi

[hasty pecan]

I need help of everyone

[languid harness]

Hello OpenAI Community and Support Team,

I am writing to seek assistance regarding the deactivation of my account (Email: [[my email]]). I have already submitted an appeal through the official channel, but it was unfortunately upheld. I am reaching out here in hopes of a manual review, as I believe my case may be a false positive triggered by my specific field of research.

About My Use Case:
I am a Professor at the South China University of Technology (SCUT), specializing in biomaterial design and cancer therapy. My work is strictly medicinal and therapeutic. For your reference, you may verify my credentials here:

Google Scholar: Xu Shidang (10,000+ citations, h-index 46).

Recently, I used ChatGPT to research tumor-specific promoters and Cis-Regulatory Elements (CRE) design. This is a standard and vital area of research in cancer gene therapy, aimed at killing cancer cells while sparing healthy tissue.

During my search, I encountered several content restrictions. Because this is legitimate medical research, I mistakenly assumed the restrictions were a technical error and attempted to rephrase my queries to obtain the necessary scientific data. I now understand that these queries may have inadvertently triggered safety filters related to biological agents. I assure you that my intentions were purely academic and focused on oncology.

Details:

Case ID: C-U7Ha6yvQcrEU

Primary Use: Academic Research & Teaching

I have been a loyal user and deeply value the role of OpenAI’s tools in my scientific workflow.This account is essential to my work. Could a team member please take a closer look at the context of my research? I am more than happy to provide institutional verification or further details about my project if needed.

Best regards,

Xu Shidang
Professor
South China University of Technology (SCUT)

[wintry parcel]

Hi, there. The staff of this server are mostly community member volunteers and can't help with OAI account issues. I think your best option is to go through OAI's help center with your questions, as recommended in OAI's article. https://help.openai.com/en/articles/10562188-why-was-my-openai-account-deactivated

[trail cairn]

you should probably quickly backup everything you can if you actually contacted the official support and said you are using it from china as it may result in account ban https://help.openai.com/en/articles/7947663-chatgpt-supported-countries

[celest crater]

if i want make video like this or use this and just change objects and keep more or same sequence is it possibile?
anyone has any tip how to do it?
lets say i want have toys instead veggies

[green summit]

What I did was:

Downloaded your recording > took screenshots of the game > went to GPT and added the screenshots + context about the kind of game it is > it then created a prompt for it.

Basically, adding images and information about the kind of thing you want helps a lot.

[languid harness]

thank you for ur reply and kind reminder!

[languid harness]

thanks! I will to it

[golden crypt]

The support email is [e-mail] if you're looking for it 🙂

https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[wintry path]

I know that the business plan is a minimum of 2 seats. If I do pay for the two seats - can I just use it with my two emails - because I do need the extra usage with the codex and pro rate limits etc - but I dont want to do the Pro sub as that unnecessary waste for me.

Is this somehow forbidden in the terms of agreement/does this fall under "abuse of rate limits/circumvention"?

I dont seem to find any clear guidelines regarding this - but I really want the 30 Pro-model requests per month + 2x codex limits

I think people here have been using it like that - I wanna know if there are any specific guidelines/has someone gotten banned for this?

Also really like the flexible pricing structure so can solo users using business accounts chime in here?

[onyx kraken]

I recommend reaching out to OpenAI tech support directly and discussing your interest and asking if it is permitted.

A guide to reaching out to tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support

The reason why I recommend, literally ask them for permission, the Business plan helpfiles state,

Business workspaces are designed for consistent, collaborative use within an organization. Misuse of seat assignments in violation of Services Agreement and may lead to workspace deactivation or account suspension. https://help.openai.com/en/articles/8542115-chatgpt-business-faq

The ChatGPT Business subscription requires a minimum of 2 users. https://help.openai.com/en/articles/8792536-manage-billing-on-the-chatgpt-business-subscription-plan

They don't say 'minimum of two seats' - and it's possible they care about not letting 2+ people share the same seat, but may not mind 1 person using 2+ seats - but they might. Those helpfiles are pretty specificly worded, so I absolutely would check and clarify to make sure it's inside intended use.

[wintry path]

Ah I see - I do see a lot of people on here using 2 seats while being a solo user - I guess I'll just have to email them to know

[wintry path]

However the support on their page is AI support right? I dont really see a place to contact a human

[wintry path]

Okay was able to get through to a human by asking "Escalate to human" - and got an email which was a non-help, but trying to ask clarifying questions seemed to help.

[glacial aurora]

I'm a fairly new and casual user, I'm trying to use ChatGPT5.2 as an assistant for roleplaying games. How can I get it to actually read documents that I provide? Often times I will provide it a PDF with instructions to read it and then ask questions about that PDF and it gives me blatantly incorrect answers. I will correct it and tell it to read the PDF and it will confirm that it has, and then continue to give me blatantly incorrect answers that are easily verified with the document. Extremely frustrating.

[onyx kraken]

I usually see issues like that when I have given conflicting instructions to the model, or something about the instructions or the material I uploaded is conflicting with its programming and/or training.

This can also happen when the model's being given more information than it can handle at once.

A way to diagnose, maybe in a new chat break it down into smaller steps and confirm the pieces. You may find which parts or that the total volume of the whole is the issue.

[onyx kraken]

I would probably ask, "Can I be the one human using each seat of a Business membership?" to try and guide clearly to the answer I am looking for.

[glacial aurora]

This can also happen when the model's being given more information than it can handle at once.
Likely this part since some of the PDFs are rpg books 2-300 pages long. So no way to work around that? Instruct it to read a specific page count maybe? IE Read up to page 100, read page 101 to 200?

[onyx kraken]

Two things to consider in this; 1 is potential copyright issues. Be aware, this could come back to bite you, because of ToS,

Your content. You may provide input to the Services (“Input”), and receive output from the Services based on the Input (“Output”). Input and Output are collectively “Content.” You are responsible for Content, including ensuring that it does not violate any applicable law or these Terms. You represent and warrant that you have all rights, licenses, and permissions needed to provide Input to our Services.```

https://openai.com/policies/row-terms-of-use/

The model's trained to try and support appropriate behavior with copyright too, and that may be part of its refusal and creating original content instead of following what you ask for - we see this with some requests involving other IP too.

About 'but just read page 20!', really does not work, because the model reads it all, then evaluates, then if successful and not overwhelmed narrows to just page 20 (or what it understands to be that page - which might not be properly narrowed to page 20 - the model doesn't really 'see pages' like we do when we're looking at this stuff).

Presuming you're using material you have the rights to, one thing you can do is discuss chunks of it with the model, pull out the core stuff that is needed to be followed (A human DM rarely uses every page, but does use core tables frequently; the model has a lot of training on various things, so it may not need some percent of the material - you can potentially pull out what the model needs, make sure it's not an overwhelming volume and there's no conflicts, and it would then likely work smoothly.

[glacial aurora]

There's should be no issues for copyright, I'm using it as a personal assistant for a personal game, no content creation or anything else happening there but I appreciate the callout. Frustrating that there really isn't a solution though. Thanks for the information!

[onyx kraken]

I would focus then on ways to consolidate the information into ruleset lengths it can handle, and check carefully for contradictions, because if there are any - or even some strong ambiguities - that can make the model appear to be not following instructions; the conflict seems to kick it off course.

[jaunty axle]

Is this the right place to ask questions about OpenAI Agent SDK?

[onyx kraken]

Sure! You may find more folks able to answer the questions in #dev-chat and/or #1037561751362863144, but it would be fine to ask here too.

[shell wraith]

Is it me alone or ChatGPT is really freezing/lagging in Chrome?

[twilit bough]

Does anyone know if there is a reliable way to force a chat into a remount? I am running into repeatable issue where an ongoing chat is unable to see new uploads in the project files while a new chat in the same project can see them fine. It seems to resolve itself after a few hours but it's unreliable and.... sub-optimal to say the least

I tried:

1. refreshing page
2. clearing cache
3. branching chat
4. asking for re-index without relying on memory
5. crying
6. begging

It sometimes starts showing new files after a while, sometimes not, I don't know if it's anything I'm doing that's fixing it or if a remount happens on a timer beyond my control, I am completely lost on this

[ember wing]

video editor

[rustic marlin]

usually, the best solution to problems with AI context is to start again on a blank chat

in the case of projects, create a new proejct and copypaste everything relevant

I know it sucks to do that, Im not saying how it should be, Im just saying that this is the most reliable way to get back to work

btw, depending on what your proejct is, consider using something like codex and codex-cli, if it is about code

[twilit bough]

The issue isn't with the context itself just with how the chats access the files. For example:

Project A, you upload a file TestA.txt, start ChatA and ask it to index the files in project folder. It will correctly say TestA.txt is in project folder. But when you then go back and upload TestB.txt to the project folder, ChatA will often not see it for up to a few hours. It eventually sees it but I have no clue what triggers the remount. At the same time starting ChatB in ProjectA creates a fresh mount and that new chat can see both TestA.txt and TestB.txt right away. I just don't know how to force ChatA to actually access the files in project folder as it exists instead of a stale snapshot of it

[rustic marlin]

hmmm

try to upload it on the project rather than on the chat

just a guess

[twilit bough]

no that's exactly what I'm doing 😄

[rustic marlin]

hmmmmmmmm

[twilit bough]

if I upload to chat then it's only visible for that chat

[rustic marlin]

im not very sure of how to help with that, tbh, I barely use projects, I often use codex

and I usually don't have problems with that, it has some rpetty decent tooling for exploring files

what is you proejct about?

[twilit bough]

I have one hobby game dev project where I'm keeping a lot of game design documents there, update them, discuss, do game stuies and such, and one other where it's a relatively simple Excell VBA workbook development for my day job. I noticed the issue on both but it's way more of an issue with the VBA project because the iterations are pretty quick. Though in general I notied that the chats seem to be incapable of distinguising between
A) files in project folder
B) chat attachments
C) files they produced and linked for download
They are all listed as /mnt/data contents and the chats are incapable of telling which is A B or C (only from context, but if the name is the same they just appear as identical files with identical directory)

It seems to me like an oversight in design tbh

[rustic marlin]

right, since it is code, really, give it a try on codex cli

the web codex is cool, but I think it is too mych "vibe-code" for my taste, takes too much of my control away, while cli lets me be more i ncharge of how things goes

for your case, the advantage of codex-cli would be that you can create a fine named agents.md and the AI will use that as a system prompt, so, you can pack all the instructions on it about yoru project structure

for example, on my most recent proejct, I have a few lines indicating there where each kind of file goes..
here is a piece of it:

- `app/` hosts the Nuxt 4 client code: `components/` for reusable UI, `pages/` for routed views, `layouts/` for shells, and `Repositories/` for data access helpers and websocket emitters.
- `app/plugins/` and `middleware/` register global behaviors; `locales/` centralizes i18n strings to keep translations close to the UI.

[twilit bough]

I'll check that out later, though im still hoping someone will have a solution for the stale mount issue ^^

Thanks

[rustic marlin]

ah, sure
sorry for not being able to help with that, tho xD

but yea, I often suggest people to often jsut migrate to more adequate tools, after a while, chatgpt feels limited, not because the AI is incapable, but jsut because of the surounding tooling

[twilit bough]

that's fair, my project has a lot more concept work and design than just code but I'll definitely give it a fair shake

[shy trail]

Hello i want to ask something, when i logged into Sora 2 it says that my session expired. Tell me how to fix that? Ive been looking at articles and nothing contributes in my problem

[lethal cairn]

Quick question: how do users get permission to post in #1070006915414900886?
I see others posting there, but I don’t have posting access yet and didn’t find instructions in the pinned messages.
Just want to make sure I’m following the right process.

[timber shadow]

Hey! There is a button to submit a bug report in the pinned message you mentioned. Have you tried clicking that button?

[lethal cairn]

Thanks Solbus! Let me clarify. I want to report a bug in ChatGPT, not Discord. I know it's a noob question, but that I certainly am. LOL

[timber shadow]

You can definitely post a bug related to ChatGPT in #1070006915414900886 using the "Report a Bug" button in the pinned post in that channel. If you'd like to reach out to OpenAI support, you can do that via the chat button on help.openai.com

[lethal cairn]

The issue, I think, is that my "verified" role hasn't been added yet (learned it can take up to an hour). I wasn't allowed to make posts in the channel.

[wintry parcel]

Did you use the "Start Here" button (not "New Post")? It'll open the post where you will find the button to open a "Report a Bug" form that Solbus mentioned.

[rugged yoke]

hi

[graceful abyss]

👋

[sudden compass]

Why does GPT lose connection around the 13-minute mark when thinking?

[strange minnow]

The connection drops because the system isn’t designed to hold a single silent request open indefinitely, and ~13 minutes lines up perfectly with common timeout thresholds.

[royal axle]

Can anyone help me out with an invite code.?

[timber shadow]

Hey! Try checking in the #sora-2-codes channel

[royal axle]

Thanks just saw it.

[ornate galleon]

The image generation seems to have massive failures when asked to create images with specific shapes I.e. create me a grid 4 rows by 6 columns it defaults to 5x5 even when specifically told dont to this 4x5 or 5x5 make sure there are only this many circles, even when told specially use “engineering mode” “make a diagram” then later when told to make 7x4 rows of circles. Why is this happening. I can have it make me a yarn version of the sun, but basic mathematic sketches it seems to lose its mind. What am I doing wrong?

[broken pebble]

Hey guys. Was looking for help with prompts for image generation. I used to use sora image gen and chatgpt image gen to make locations for my game, but ever since they changed the model used to make images, the outputs are consistently rubbish.

I generally do this for my prompts:

[Intro block of text that is ALWAYS THE SAME between images]
[The location is: Description/vibes of the location]
[Re-iteration of art style, this is ALWAYS THE SAME between images]

In the past, this has made plenty of nice looking images, and i could change the middle section to basically whatever and it would have an extremely similar art style.

With the change, the old prompts are destroyed and abhorrent, and every attempt i've made to create a new prompt base has failed; i still havent worked out how to get consistent art style images yet

this is the type of images i used to generate and the prompt used for it. You can try the prompt yourself and youll see that the outputs are pretty digusting and bad

I'm not sure how to describe the art style;
this is the closest i got with a new prompt but this is soooo inconsistent.

[steady garnet]

example of the results you don't like?

I'd start by using a specific style declaration like "in the style of 1780s explorer paintings" instead of the very vague "masterful depiction" or "immersive fantasy"

[broken pebble]

new on right

(same prompts)

to be fair the ones on the left are the best out of like 30+ each but
ALL the new ones look like the ones on right

[steady garnet]

You might want to give directions on the saturation, maybe emphasize the anime part by referencing specific artists or styles. I think it did pretty well in the top right but didn't get the same color scheme in its tendency to overdo texture and detail. You might specify less transition and shading

Yeah the island ones its making for me are really bad. I guess the model is complex enough that more specific art technique awareness could be required for a specific look over a specific view. But I think they probably just need to re-release 4o image generator because a lot of people like it better

so changing to 1780s explorer painting got the color pallet better for me

make sure to do any changes out of a fresh conversation

5.2 is very context dependent so it'll just keep running the same ugly ones over and over

[wintry parcel]

Hi, there. I've found the current image model (GPT Image 1.5) to be very good at photorealism but not so much with painterly styles, and it takes a lot more detailed prompts to direct the art style. Also　the last model (GPT Image 1) is still available through OAI's custom GPT https://chatgpt.com/g/g-6940a876d5f4819186b4668deabcd580-4o-imagegen.

You might also post your question in https://discord.com/channels/974519864045756446/1060915255720558592

[steady garnet]

huh, I can't paste my images into images discussions for some reason

[wintry parcel]

Correct. It's a text only channel for discussion. You can share your images in https://discord.com/channels/974519864045756446/1204360881593520128

[rugged fern]

chatgpt is always laggy and even chrome says "chatgpt isnt responding" with options to wait and close it, is it a chrome issue or chatgpt issue? the other info is that the chat is very long, ive been using it for a lot of time, its a big chat. but i hate it gets stuck any ways to fix?

[charred parcel]

Try the ChatGPT Desktop App rather than a browser

[rugged fern]

same issue I guess

[charred parcel]

on macOS it's fine with long chats

[rugged fern]

windows user here

[charred parcel]

have you tried the ChatGPT Windows Desktop App?

[rugged fern]

yes

[charred parcel]

no good?

[rugged fern]

nope, still same issue. laggy. but if i start a new chat everything's very smooth

[charred parcel]

ok well then I guess the answer is to have shorter chats

[rugged fern]

since you said it works fine with longer chats on macos, is there any way for windows?

[charred parcel]

the ChatGPT Desktop App for Windows is just a wrapper as I understand it, so there's not much you can do. The macOS app is native and doesn't suffer from browser limitations.

[rugged fern]

damn 😭

[mighty garden]

On phone even long chats work fine... For me..

[delicate onyx]

Hello Everyone,
Please, I'm looking to host OpenAI workshops and meetups at my university. Who should I communicate with?

[onyx iron]

I hit my weekly codex limit. I paid $40 to increase it. Nothing changed. I did get 1000 code review credits but i do not want that i want my weekly limit lifted.

I asked for refund but it keeps saying to use apple to process but i did not purchase through apple!

Can anyone from open ai help here?

[toxic kelp]

Please approve my waitlist for gpt health i am a cancer patient and will use my data to help the model

[keen haven]

Anytime i hit the join the waitlist button for gpt health, it goes to a 404 not found page, any help?

[naive atlas]

Same here

[onyx kraken]

I would reach out to/through https://forum.openai.com/home , this Discord is OpenAI's Discord, but we're primarily community members here talking to each other online. The forum is a bit more directly tuned towards OpenAI talking with people, at least during fairly-regular scheduled online presentations, and there's membership groups and a lot of other university members there from a great many places.

Be here too! But we're less set up for things related to hosting workshops and meetups, that's maybe closer to how the Forum is working.

[onyx kraken]

We're almost entirely community members, like you!

https://help.openai.com/en/articles/6614161-how-can-i-contact-support for tech support, including refund support

[onyx kraken]

Checking, per https://openai.com/index/introducing-chatgpt-health/

"We’re starting by providing access to a small group of early users to learn and continue refining the experience—users with ChatGPT Free, Go, Plus, and Pro plans outside of the European Economic Area, Switzerland, and the United Kingdom are eligible. As we make improvements, we plan to expand access and make Health available to all users on web and iOS in the coming weeks."

Are you a member with a supported plan, in a supported region?

[toxic kelp]

Lame

[keen haven]

Yeah i should fit those requirements, i have chatGPT free and im in the US

[onyx kraken]

I'd check for a system problem them, if you can connect on another device, or another network. If you need to use the same system, consider a reboot, clearing cache, if there could be an issue with browser extensions (usually need to be removed, not just turned off) or VPN (can usually be turned off).

[onyx kraken]

Or not! I checked and I'm getting a 404 as well, though I haven't seen anyone else report that.

[keen haven]

Yeah i’ll try again in an hour or something. I checked on my laptop and it gave the same thing again.

[onyx kraken]

So sorry. Hopefully it works soon, starting to see a few reports in other channels too

[keen haven]

No problem! Thank you!

[long crag]

Is the Health waitlist instant?

[main phoenix]

Hi - I don’t see connector (ChatGPT Apps) option being available in custom GPT’s but I do see it in general chats. Is anyone else experiencing this issue. What are ways I can get access to apps/connectors in my customGPT?

[glacial aurora]

Is there a way to get ChatGPT to not just blatantly lie to you? I asked it if it could create a short video file with sound and it said yes. When I gave it a prompt to do that it clarified that it can't actually do that. So I asked if could create a gif of the same prompt and it said yes, then when I told it to do so it said nope, can't do that either... like, what was the point of that?

[soft gate]

Why is it whenever i try to make a video it says “update your app” and i did go update it but this issue proceeds to keep happening

[civic stratus]

is there a way to duplicate a conversation on GPT?

Let me explain - I am generating a bunch of art images. I have talked with GPT and got it to where I would like. Now i want to generate a bunch of images in multiple tabs with the current information we have had from our chat thread.

Essentially would like to duplicate GPT conversations and have it be unique (one of kind) conversation using the data from that point

[wintry parcel]

I don't know if we can "duplicate" an entire convo but if you are trying to recreate a conversational context to generate a series of images in that particular style across multiple chats. you can try the following:

1. In the original chat, ask ChatGPT to create a detailed style guide based on the entire conversation as a downloadable file (like txt or docx). 2) Review the style guide to make sure all the elements, are there. 3) Open a new chat. 4) Upload the style guide and a few sample images, and tell ChatGPT that you want to generate a series of images in the exact style defined in the style guide and uploaded reference images. Ask ChatGPT to review attachments and acknowledge the goals of this convo. 5) Provide an image prompt once ChatGPT acknowledges.

[proud moon]

can chatgpt get access to a private repo from github if the connection is enabled in the console settings? It always tells me it cannot read my private repos 🙁

[fallen ravine]

How do you fix ChatGPT being unable to see an uploaded file

Specifically CSV

[wintry parcel]

I got curious about your problem and had a long discussion with ChatGPT. Long story short, asking if it's able to generate a video is not specific enough. It said "no" when I asked whether it is able to generate a video in the chat.

[glacial aurora]

That's kind of crazy, I deleted the conversation to avoid clutter so I can't clip the response but IIRC I asked it:
Are you able to generate an animated gif? (Obviously responses are paraphrased)
It responded : Yes, blah blah blah
I said: Actually, are you able to generate a short video clip with sound?
It responded: Yes I can, up to 1 minute with sound blah blah blah. I can generate it based on an existing clip, extrapolating from an image, from a story board or even just from a description. just provide the direction and then tell me you would like to generate an .MP4 with sound!
I gave it a description of the clip I wanted and it responds: Sorry, I can't actually do that [chatter about tools that exist to create videos].
Me: Okay, please create an animated gif based on that description.
GPT: Sorry, I can't do that. [Chatter about Gif creation tools].

I am fairly positive it was a fresh chat with no history and not part of a project that could have caused it to inherit other instructions or context.

[wintry parcel]

Generally speaking, ChatGPT (and many other AI chatbots) don't do well when it comes to recognizing their own internal workings. ChatGPT gave me a long-winded explanation about how it got confused between video-generation anywhere vs. right inside the conversation (basically). That sounds crazy to us humans but we have to keep in mind AI thinks differently.

[glacial aurora]

You're not wrong, but when you purchase a subsription to a service, it becomes a very different conversation. As a paying customer, does the product do what it's supposed to do? We've all gotten far too comfortable 'understanding' flaws when companies should be held accountable for minimum standards of quality. Regardless of technical limitations or it 'being understandable' a paid product blatantly answering simple questions wrong is not something we should tolerate as consumers.

[tawdry ledge]

hello, I'm trying to reduce the seatcount of my ChatGPT business plan from 5 to 3 seats before our plan renews in a few days, i've removed 2 of our members and we have 2 seats free but it's not letting me go below 5 seats. Is there any place to contact for help?

[desert lark]

Is there something going on with chat I keep getting this error and I am using a gpt I made and it use to work fine . Here is an link to it https://chatgpt.com/share/695fff40-0018-800d-834a-a5e0f2f2ecc8 I even asked to fix my gpt prompt and i remade a gpt and it still gives me the error.

[neat hemlock]

If your goal is to create a similar video, it depends on whether you want it fully AI-generated or edited from real footage.
For AI-first workflows, tools like Runway or Pika are good for short, stylized clips. If it’s more about control and quality, a hybrid approach works well: generate scenes with AI, then refine timing and transitions in something like Premiere or DaVinci Resolve.
If you can share what part you want to replicate most (style, motion, or realism), it’ll be easier to suggest a clean setup.

[neat hemlock]

I am really happy to join in this channel.
@wintry path
From what I’ve seen, most teams start on the business plan mainly for reliability and higher limits, not because it’s perfect. The AI-based support can be frustrating, but once usage stabilizes, the value usually comes from predictable throughput and fewer rate-limit surprises.
For early-stage products, I’ve noticed people succeed by tightly controlling prompts, caching aggressively, and only scaling model usage once user behavior is clear.

[neat hemlock]

Strong advice. I’ve seen too many cases where people assume exports or history are permanent until something changes upstream.
A simple local + cloud backup (even a weekly one) already removes most of the risk, especially when experimenting with new tools or betas.

[smoky shore]

is there anyone who can tell where i can generate sora 2 video free or daily credit limit

[north shadow]

every single time i try to send an image to get help with, it now gives me this. Why? What's the damn problem suddenly? this is not acceptable.

and when i retry it delets my message entirely and acts as if i did not send one.

[dusk crescent]

Do the app connections only work on gpt5? I have the free plan, and the apps seem to stop working when I run out of chats

[vernal wyvern]

I can make chat-gpt search the internet for anybody i want to, including email address, phone numbers, anything I ask gpt will do. This is a huge problem

[merry minnow]

help

[wintry parcel]

"Simply leave those seats unclaimed and we'll adjust your invoice at the end of the billing period, taking into account any unused licenses." Quoted from https://help.openai.com/en/articles/8792536-manage-billing-on-the-chatgpt-business-subscription-plan

[quartz spade]

can someone tell me is the sora image generator the same model/quality as the chatgpt one? i cant find any model info

[charred parcel]

they're the same

[tawdry ledge]

oh ok

[tawdry ledge]

thank you!

[terse sand]

Anyone having issues with 5.2 Pro not showing its chain of thought in real time? I'm on an enterprise account if that makes a difference

[glacial aurora]

Is it no longer possible to branch a chat?

[ripe lily]

im having it do this, im not sure if related?

[gaunt jacinth]

Hello everyone,

I have a question about the Milestone Awards.

We passed the 10B level on August 31, 2025, but it looks like we missed the reward because the first batch cutoff was July 2025.

Since then, we’ve also passed 20B tokens, but we still haven’t heard anything from OpenAI. Was there supposed to be another batch or a new round of awards?

I’m seeing people on LinkedIn still receiving Milestone Awards regularly, so I wanted to check if there’s an updated process or anything we should do on our side.

Thanks!

[rugged fern]

ye, on phone longer chats work somehow

[vernal wyvern]

Guess I'll just report it as ignored and sent it to the EU.

[wintry parcel]

Hi, there. Have you contacted OAI customer support? More info at https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[ivory hare]

anyone good at removing background noise

[vivid mortar]

From audio or from images?

[simple geyser]

Did chatgpt heath release already?????

[timber shadow]

Hello! If you are interested, check out the announcement here: #announcements message In the link, you'll find the button to join the waitlist.

[copper sun]

How can I delete conversations starting from a specific message in chatgpt? If I ask it, will it delete them automatically, or is that not possible?

[copper sun]

I don't want to delete the entire chat conversation.

[green summit]

You can't really "delete" messages, you'd have to "edit" the message you sent, and it'll reset the chat from that point onwards. It will not remember the conversation that got "deleted" after editing the message.

Since I can send videos here, I'll screen record and show you. One sec

[green summit]

@copper sun

Hope that helps!

[green summit]

"My gut" says black, but white always looks nice, too. XD (not an answer to your question, but yeah 😌)

[green lily]

eh

no one saw that

so uh wat is this all about as of now?:

black

[fast rover]

Hi everyone, I need help. My API account was deactivated for "weapons" policy violation, but I never used the API for anything related to weapons. I'm a developer working with chatbots and automation.
I tried to reply to the deactivation email as instructed, but it bounced back (delivery failure). I already opened a support ticket but need this resolved urgently.
Case ID: C-nqivfunBZTDg
Organization ID: org-ogVVDp6ln9tzZ2Qn5BEptm3s
Has anyone experienced this? Any tips on how to get a faster response from OpenAI support?

[long shore]

Hi everyone,
I’m a ChatGPT Plus subscriber, but I’ve been unable to access GPT-5.2 on any of my clients (web, desktop, mobile). Has anyone else encountered this issue? I’ve tried clearing caches, logging out and back in, but I still can’t switch to GPT-5.2.
If anyone has experienced this or knows how to resolve it, I’d really appreciate your help. Thank you! 🙏

[carmine finch]

hey guys question
i use chatgpt free with 10 messages a day limit
how many messages a day can i send with chatgpt go?
should i get go or plus?

[subtle nexus]

i have a problem

Hi, my Google account was deleted from my ChatGPT account, where I had an active Plus subscription, and now I can't access it. I tried changing the Gmail address, but that account is connected to this Discord account.

[stoic fiber]

same problem here actually, i'm wondering if it's because of a failed subscription

well similar, can acces gpt-5.2 but i can't access codex, keep getting a rate limit error even htough i haven't used the service all day

[crystal steppe]

Quick question about chat gpt?

[teal nest]

rate limits are dynamic and so there is a chance that due to high traffic it is extremely low for plus users at the moment

remember: enterprise>pro>buisness>edu>plus>go>free

[stoic fiber]

Hmm

[teal nest]

are you also sure you haven’t hit your 5h or weekly codex maxes? I seem to remember that they are quite limited on plus

[stoic fiber]

Oh, maybe thats it

Ill wait until the timer runs out

And ask again

[crystal steppe]

I'm using chat GPT Plus to make a very large document I've kept it all in one thread and now I'm asking it to go back and create the document and it says it can't do all of it at once?

[teal nest]

you can check on the codex site online in settings or if you click on local in the ide (assuming you’re using that) and then rate limits

[teal nest]

copy paste the whole conversation cmd a then cmd c on mac into a word docx then have gpt thinking index the parts and then have it fill in the context in a new chat feeding it both results and then asking it to reply with a completed docx, after indexing you may have to prompt it to fill in section by section in that case make sure that it is sending the running docx as it works and remember to use thinking

[crystal steppe]

I'll try that out thank you

[teal nest]

in order to do that to copy it all you do have to scroll all the way up and load it all first though

[heavy patio]

Did you use the pro model to ask?

In my experience, the Pro model is really good at making documents. The other models have been sporadic on their success at that particular task, for me.

[wintry parcel]

Hi, there. I'm not sure if I'm understanding your situation. Are you using your Google account to sign into ChatGPT? And how do you know your Gmail account was "deleted" from your ChatGPT account?

[teal nest]

they are a plus user

[heavy patio]

Lol, sorry. I totally missed that. Read too fast 🙂

[teal nest]

but that’s what i would do too

if they were pro

that’s why I said section by section after indexing to make up for it not being a pro model

so it retains detail

[sweet widget]

Hi!

I have this situation an i think someone can help me

I have a landing page for my company that's completely finished. I created it with HTML, CSS, and JS in Visual Studio Code, along with the help of Claude Opus 4.5 AI extension. My landing page is supposed to have a custom AI about the company using an OpenAI API key, but it's not working at all. I don't know how to make my custom AI work. Could you tell me what other data I need or what I should do to get it working

[past storm]

anyway to contact a human in support i want a refund i forgot to cancel my subscription and well got charged and i would like an refund

[harsh obsidian]

Hey! Hey! We can’t handle billing or refunds on Discord, but you can learn more about refunds here: https://help.openai.com/en/articles/7232895-how-do-i-request-a-refund-for-my-chatgpt-subscription.

[crystal steppe]

I hope so, it doesn't specify it says chatgpt 5.2

[teal nest]

is your front end tied to your backend?

[fast arrow]

Hey my request character has a problem

[rustic marlin]

you may want to ask more on #dev-chat

[indigo harness]

It's been 12 days into January nd I still am not able to access the Year with Chat Gpt feature
It has already rolled out to most Indian accs(I'm in India) but I still cannot access mine

Any1 has a solution for it?

I've been pretty active with chat gpt this year nd hve the reference saved memories features also turned on btw

[zenith solstice]

Hi! Any Therapist GPT (or other AI) recommended? Just wannat try different bot and see what's the difference between them, many thanks!

[white meteor]

Does open ai have a discord ai chatgpt bot?

and is something like that ever going to happen?

[naive bramble]

bro when is sora 2 coming out for other countries other than the usa. like we've been waiting for soo long

[wintry parcel]

There's no published schedule for future rollout.
Plse keep an eye on https://discord.com/channels/974519864045756446/977259063052234752.

[heavy drum]

Hi, sure.

Dm:d you 🙂

[heavy drum]

I’d recommend Robin Life Coach and CounselorGPT. I would also recommend Attachment Theory Therapist.

[lunar magnet]

is there anyone in here with pro that could answer some questions on the differences between plus and pro?

[rocky seal]

SORA will not finish generating my video. Been waiting 20 minutes

[minor bolt]

So a object is going to be in the middle of my video
Not in starting frame
How do I tell sora to use this particular object??

This truck is approaching from horizon
So its not in starting frame

[zenith solstice]

Thanks!

[green summit]

Unfortunately, that is not possible. You'd have to have one of the AI make an image with the truck already on the horizon while still fully visible when it comes to the markings on the truck. Which is gonna be very hard.

[minor bolt]

How did you create the 3 minutes video on your YouTube?

I get only 15 seconds

[green summit]

I put many clips together using the Samsung Gallery's Studio mode.

[minor bolt]

I am trying that but the transition is not very smooth
I am using the last clips last frame as first frame of the next clip

[green summit]

I also used first and final frames. Idk if you use a tool to remove the watermark from the image, but if you do, and it's "Magic Eraser" then I understand why it doesn't transition smoothly.

Or what exactly do you mean with smoothly?

[minor bolt]

like i had a clip before this and after the scene

i told him to transition and it gave this :

[green summit]

What generation tool do you use? Is that Grok?

[minor bolt]

social sight ai

seedance 1.5 pro

[green summit]

Ah, I see

[minor bolt]

it was only one who allowed two refernce images

one of starting adn other of ending

i also tried veo but i tell to make video thrilling and fast but it always makes slow

[green summit]

Minimax/Hailou does too, I believe, but idk if that's good for what you want. It's mainly for animation if I'm not mistaken.

[minor bolt]

VEO 3.1

[green summit]

Do you use any LLMs for your prompts?

[minor bolt]

these were outputs of sora 2

[minor bolt]

yes

i give chatgpt the cookbook link and tell him to analyze

[green summit]

Yeah with Sora it's basically impossible to get it how you want without losing the markings, etc.

[green summit]

Alright cool

[minor bolt]

not completely, i describe the scene in detailed but not so good in english so i just tell him to format properly using cookbook

[green summit]

Yeah it's usually good (from my experience at least) in knowing what you want and have it make a prompt for it.

(Usually)

But with things like the logos, well, it's possible but a lot harder to incorporate into the prompt.

[minor bolt]

Not only logos , it messes up the camera movement too

[minor bolt]

In both of these videos i told him to smoothly move the camera 180 degrees as the trucks pass by

But in one video the convoy got longer than video length and in other idk what the hell he was thinking

[minor bolt]

This video would have turned soo good but idk what happened in between
It decided to give me drone of 6 lane highway

I can still cut the video

[karmic aspen]

If anybody else experienced similar issues and got it resolved, please let me know how to deal with it.
Thanks in advance.

[eternal meteor]

Does somebody know a super free voice cover ai for a song?

[dense stag]

Hi , does anyone have suggestions for the following challenge:

i am continuously getting “suspicious activity detected”. Followed all instructions on hep page (reset password, added mfa, and logged out of account on all devices. i don’t use proxy or vpn. i’m not sharing my account with anyone. have waited hours for it to stop the flag. and support isn’t able to do anything.

anyone experience this before?

[hoary quarry]

Hello Banchot!

[upper pike]

Is anyone teach me prompt engineering?

[onyx kraken]

I consider this to be the core of prompt engineering:

1. pick any language you know really well that the AI understands too.

2. understand exactly what you want the AI to provide.

3. explain this, focusing on what you want the AI to actually do. Using language as accurately as you can, avoid typos and grammar mistakes and communicate clearly as possible.

4. check the output carefully, verify you get what you intended. Remember to fact check, and be extra careful with any math, sources, code, or other details that the AI is known to be especially likely to hallucinate.

[simple geyser]

Is gpt health on android now???

[heavy inlet]

Hi, I’m a ChatGPT Pro user.
I run an e-commerce brand and am actively producing product and brand videos.
I’m very interested in using Sora for cinematic product storytelling and commercial content.
Happy to provide feedback and real-world use cases.

[pallid folio]

Why did ChatGPT on Android suddenly force me to subscribe to GPT Go and not let me see other plans like Plus or Pro?

[glad heath]

Is chatgpt slow since chat 5.2 update or it is just me ?

[smoky sluice]

My Plus subscription was renewed on December 21st. The system showed the expiration date as January 17th.
The plan was active on January 13th.
On January 14th, the account became free earlier than the stated date.
Please help

[verbal crag]

Hello! You may contact OpenAI support through this link: https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[ionic crescent]

my search engine is so far to the left i cant do anything... and the same is true forsome other windows, and yes i tried zooming in and zooming out... chatgpt doesnt help either

[woven tree]

Can someone help? Before answering: yes, I’m from the USA, and yes, I didn’t use a VPN.

I bought Sora Pro, but video generation was not available at the time. After some time, Sora started generating five videos for me and then got stuck on the loading screen for a long time. I waited for about 30 minutes, but I was in a rush because I had to upload a video to my channel.

I saw an article saying that deleting data might help. My plan was to delete my Sora data and then regenerate a single video. However, after I deleted my data, my entire account was deleted, with no option to recover it.

[oblique bear]

I have a problem with the image generator. Most of the images I generated aren't loading which includes the images I sent to the chat. I tried everything to fix it. Went on my phone, went to a different browser, reset the site settings, everything. Nothing worked. Is there a problem?

[tropic juniper]

I'm not sure if this chat is the right place to ask but, how long does it take to be verified for Individual?

[drifting hinge]

Someone familiar with this

[lethal laurel]

can anyone tell me why every other AI is completely useless when it comes to math but chatgpt and it has a image limit so I can barley even use it

[teal nest]

what other system are you trying?

[lethal laurel]

gauth, grok and snap ai

[teal nest]

both gauth and snap ai rely on the big providers, openai, google, anthropic etc for their models, however because their models are free to use and do not use dynamic routing similar to that of when you use these models directly from the providers themselves the ones you interact with on these sites are actually even lower than the low models already provided for free by the providers themselves. Nothing is free, it’s just designed to get you to pay or to sell your data, and that’s why direct usage of models will always be better; the image limit is there to get you to upgrade so they make money

[gentle bough]

yeah it seems its a thing, i got blocked as well because of "Fraudelent Activities" for nothing i have done

[sharp coral]

@lethal laurel LLMs are generally bad at math, because they aren't able to calculate anything - all they do is very sophisticated guessing what words (or token) should come next. If an AI solution does perform well in complex math, it is because there is some engineering on top to cover that use-case. And regarding the image limit: Are you uploading images of equations/texts? If yes, simply try using one of the bazillion OCR tools, which can extract text from images, and use those directly, so you won't have to worry about the limit anymore.

A general question to all: What would be the appropriate space to place a feature request for ChatGPT (user frontend)? Only if there is one at all, and if such feedback actually every reaches someone at OpenAI 😄

[charred parcel]

#1053565716500074526 is probably the best place

[sharp coral]

Completely missed it - thanks!

[lethal laurel]

Alright ty

[flat sorrel]

Greets, how do I ask chat GPT to stop lying and to check on the net before answering gibberish things, even when noted in the personalized gpt's he still hallucinates way too often

[sharp coral]

You can tell him to provide sources for all factual claims, although that will make every single response take much longer than without.

[flat sorrel]

i'll try to ask sources. time is not an issue, thx

[bitter anvil]

hi. my cards are declined. what should i do?

[elfin robin]

chatgpt is broken for some reason. tried clearing cache, changing dns, restarting browser and pc. it works on my phone with data, but it does not work on the pc while having hptspot from my phone. idk what to do

[short star]

Hi guys, my name is Jose. I work at a Gas company in Brazi,l and I'm trying to create or find an AI model that can detect the number of TARE of gas recipients. I want to find one that does not require paying for tokens to use it. I currently have 1000 pictures of tares. To accomplish my goal, what do you guys think would be the best option

[flat sorrel]

transform pictures to text with text recognition, make a list in a big text file and submit the text to any IA.

you can maybe even automate it with make or something

[wintry parcel]

Hi, there. You can also select "Web Search" or "Deep Research" from the drop down. The latter will give you a complete report with sources.

[flat sorrel]

and shall it help him stop telling lies or hallucinatins? ... I'll try thank you

[wintry parcel]

I've used Deep Research a bit and so far, all the sources checked out. As always, we need to provide a well-defined topic and clearly stipulate additional conditions/parameters to get the output close to what we want . It takes 15+ minutes to do all the work.

[shadow rampart]

Guys how long is the grace period for card payment?

[flat sorrel]

oki, thx

[rain cradle]

How do i upgrade to go

[balmy crest]

It looks like a long string made up of the repeated pattern “_aero… / _aero… / _aero…” over and over again.

There isn’t any hidden meaning in it — it just appears to be:
• either a test string,
• a glitch/encoding artifact, or
• someone pasting a fragment of corrupted text on purpose.

If this came from:
• a website,
• an app,
• a document, or
• your phone,

tell me where you saw it and I can explain why it appeared there.

When I sent a picture to ChatGPT it said that this is what it saw

[meager rune]

In the most recent announcement.. it's said that GPT Go will cost money... So my question is what would happen to the ones who subscribed to GPT Go when it was made free? Will it remain free until the subscription ends... That is... For a whole year... Or will it cost us money from the next month? Please clarify this for me... It's urgent

[modest wyvern]

I have a question, so I have ChatGPT plus, and I was wondering how to get access to GPT o4mini o4 mini-high, o3, or o1? Couldn’t find out how.

[green summit]

If you have a year plan, you'll just stay on that plan until the end date. After that you'll have to pay.

[wintry parcel]

Hi, there. Plus users have access to 4o only at this point.

[modest wyvern]

No o4-mini? I don’t care about 4o does are different models.

[green summit]

I'm Plus 😌

[wintry parcel]

This is what the OAI Help Center says: "Plus users will continue to have access to GPT-4o, while other legacy models will no longer be available. " https://help.openai.com/en/articles/7864572-what-is-the-chatgpt-model-selector

[modest wyvern]

How?

I only have 5.2 thinking and instant as well as 5.1 and 4o.

[green summit]

Idk, I just enabled legacy models in settings when GPT 5 came out and has been there ever since. But I've seen other plus users that do not have all these. So it's kinda weird ngl

[wintry parcel]

Article was updated 5 days ago, maybe the changes are still rolling out?

[green summit]

Maybe has something to do that I've used those models when they were still "relevant". Not sure exactly.

[green summit]

Hm possible, but it's odd because I've had these models in the legacy list since the release of GPT 5.

Or possibly an update later

Not confusing, like at all! 😂

Only logical explanation would be because I used the models in the past. And users that haven't won't have those in the list. But that's just speculation.

[drifting hinge]

It works for me now again, finally. But still wish you well.

[wintry parcel]

I think the plan was to phase it out after a few months for Plus users. Pro, edu, and enterprise users only at this point.

[rough herald]

is there a way to have our chatgpt stats ? like number of messages and stuff...

[green summit]

And that's still the plan, or has something changed? (I know you're not affiliated with OAI, but maybe you know)

[wintry parcel]

I'm just going by OAI announcements and articles, like the one I posted above.

[green summit]

When 2025 ended, they released something like a 2025 "rewind" feature. Which would show stats for 2025. But that's not available anymore i think. Not sure how to check otherwise.

[rough herald]

alr thx

[green summit]

Alright, I see.

[deft patrol]

My screen is saying "You've reached your limit for GPTs. Upgrade to CGPT Go or try again after 10:00PM", the little popup is stuck on my screen, and I can't get rid of it. It's pissing me off. Anyway know how to fix it? (Mobile)

[true gate]

Hey there, does anyone else noticed some quality drop for the 5.2 Thinking model in the last days?
I noticed the model started to use emojis like the normal 5.x Instant always did. It also is not really following my intention of the question, that I was after in my prompts (For example asking a question regarding Codex usage weekly limits and got an answer for the 5-hour limit). Further its sometimes answering Instantly while the model is set to 5.2 Thinking with the long thinking option on.

In the personalisation settings I found some new emotion options and that it has reset the cynical style back to default. And even after setting the new emoji drop down to less it still spams me with these random green ticks everywhere.

[balmy crest]

What was your original prompt? Never seen this type of error before

[balmy crest]

I agree tbf, since it's developed a personality it's become more open to making mistakes

[flat quartz]

Passages about prayers being hindered

[flat quartz]

It has been a recurring problem the last few days.

[balmy crest]

Try seeing if there's an update available on the AppStore? Or sometimes you can ask ChatGPT why it rejected your request

[flat quartz]

“print the 10 tenet article, give me the original” pulling from RAG document. It only happens with scripture so far. It happens with phone app and Windows.

[flat quartz]

I e asked a lot but it keeps giving different answers. No telling which is right if any.

[balmy crest]

I think there's an update available on theapp store, give that a go perhaps??

Not sure this is weird I haven't seen that error message before

[flat quartz]

It happens with custom GPT’s and blank prompts. But it never happens with prompt sent over API calls (yet)

[balmy crest]

Yh yh

Tbf my ChatGPT app today has been really bad, it can't register images. Had to temporarily use safari, it worked for me, try giving that a go

[flat quartz]

Updated. If it helps, I need to figure out how to update the web content too so it stops happening on pc. Lol

[balmy crest]

Haha cool good luck bro

[verbal warren]

Can someone from open ai assist me?

[charred parcel]

you can reach out to support at https://help.openai.com

[ornate urchin]

Thought I'd check here to see if anyone else was having problems with processing images. I'm using the Android app

[verbal warren]

I did. I've been waiting all week for a response.

[charred parcel]

go back to the help chat and ask for your issue to be escalated to a human

[verbal warren]

It asked for email.

It says escalated, and says I can add additional comments in the conversation.

Which I did yesterday.

[charred parcel]

ok good, let's see if that helps

[verbal warren]

Do I need to include something related to the issue at hand? 🤔

[charred parcel]

you need to provide as much detail as possible

[verbal warren]

How do I tell them I'm dealing personaly with my mental health and the CSAM report was because I had questions on how early exposure to such material would negatively impact the individual in question(I was exposed of my own doing at 4 y/o)?

[charred parcel]

just be as clear and succint as you can be

[verbal warren]

Thank you.

[verbal warren]

I've written out the issue with explanations. Do I leave that chat?

[charred parcel]

just leave the chat open

[verbal warren]

Okay. How long for a response most likely?

[charred parcel]

I don't know, response times vary depending upon many factors

[verbal warren]

Thank you for your time.

Response says 'we couldn't find your account. Please try again using the email address you signed up for OpenAI.

Which I dont have an account. Im trying to reactivate it. Thats part of the problem. 🥲

[charred parcel]

did you receive an email from OpenAI saying you were banned?

[verbal warren]

They said they deactivated my account yes. I had just upgraded and was hoping for a refund, or to get my account reactivated.

They said I was banned for CSAM, which doesn't make sense with what that term actually means.

[charred parcel]

that's quite serious. I doubt you will get your account reactivated for that.

[verbal warren]

Then I'd simply like a redund. I had just upgraded, and was immediately banned.

[charred parcel]

nobody on this server can help with that - the only way is to enagage with support

[verbal warren]

Which you need an account to even get noticed......

[charred parcel]

you could contact your payment services provider eg: bank, credit card company etc and ask them to chargeback the transaction

[verbal warren]

I guess that's my only option. Thank you for your time.

[wind thistle]

Hey everyone, I wanted to ask that how can I optimize my chatgpt, like its lagging a alot. It legit takes memory usage upto 1 GBs
in chrome

[covert wolf]

I have a question. Since this morning, Sora 2 has been "high throughput at the moment. Try again in a few minutes" and I can't create a video, does anyone know why?

[silent oar]

Sorry - can someone tell me off the top of their head what the Sora limits are for Plus and Business? Plus was 200 images, right? What about Business?

[silent oar]

Same limit as Plus

[green summit]

Free users:

• 30 gens/ a day day (30x 10s | 15x 15s)
• Queue at the same time: (3x)
• This might change as usual.
• (there's currently a "bug" that shows 10 gens for PC/Desktop users, but the mobile web version does show the full 30.)

Plus users:

• 30 gens/a day (30x 10s | 15x 15s)
• Queue at the same time: (3x)

Pro users:

• 99(100) gens/a day and access to both Sora 2 and the Sora 2 Pro model
• (10/15 and access to storyboard, which gives you the option to do 25 seconds.)
• Queue at the same time: (5x Sora 2 | 3x Sora 2)

[Standard 15 second clips count as 2x 10 | = 2 gens]

Additional Pro information:

• Sora 2's 25s (Standard Resolution) = 3 gens | (High Resolution) = 12 gens.
• S2 Pro's 10s (Standard Resolution) = 4 gens | (High Resolution) = 25 gens.
• S2 Pro's 15s (Standard Resolution) = 8 gens | (High Resolution) = 50 gens.

[Limited Time Only Discount for Pro users.]

• Sora 2 Pro low res 10s = 2 instead of 4
• Sora 2 Pro low res 15s = 4 instead of 8
• Sora 2 Pro high res 10s = 5 instead of 25
• Sora 2 Pro high res 15s = 10 instead of 50

[Gens now resets all at once]

"At what time does it reset?"

[Profile > Settings > Usage to see how many gens you have left | when it resets (it shows exact time and date) | and to purchase more gens]

[Purchased Gens will be available for a year, they expire after that]

Gen costs:

• 25 gens = $10
• 50 gens = $20
• 100 gens = $40
• 250 gens = $100

Final Notes:

• Using a VPN to access Sora 2 is against TOS, and might get your OpenAI account banned without the possibility of getting it back. So at your own risk ofc.

• There are no Sora 2 subscriptions, the subscription are mainly for GPT itself.

• A GPT Go subscription is available in certain countries, it's basically a budget friendly subscription. (Below Plus)

• Purchasing a Plus subscription will fix the 10 gens on PC/Desktop [yes], so the 10 gens might become permanent for free users at one point soon. (Costs = $20/ a month)

• A GPT Pro subscription costs around $200/ a month. If you're thinking about purchasing GPT Pro just for Sora 2, it's not worth it.
  [However, that last part is just my opinion and you shouldn't let my opinion influence your decision]

• and no.. this was NOT written by AI.

Whoops.. you're asking about Old Sora, aren't you? 🥲

[silent oar]

Yes, I'm a poor European 😄

[silent oar]

But thank you!!!

[summer obsidian]

is sora 2 the "use new sora" when entering the old sora tab

[atomic jolt]

Yes.

[latent hill]

Does anyone know if gpt ads will be invite/referal only

[snow steppe]

I need help for "Unable to load Conversation" error , I hope here is right place to get help. Long story short 3 days ago I randomly opened and account on Chatgpt and had some long conversation about one topic. Then I clicked to share conversation and saved the link. After 3 days I clicked on the link in different devices(differrnt mobile browser and laptop) and I receive Unable to Load Conversation error. I do not know what to do. The account I created conversation is in different laptop and in different country. I wont be able to go and check it now. Is there any solutions?

[charred parcel]

the best thing to do is contact support at https://help.openai.com

[wanton carbon]

sora now working now..?

[coral aspen]

Why have I suddenly started hitting rate limits on Plus in the past week? It timed me out for text generation for an hour despite minimal usage. Not to mention the image generation getting limited after 5-ish generations.

This is a paid subscription so thankfully it is not something I need to put up with if it continues until the 31st

A paid subscription in a competitive market lol

Is it a regional issue? Im in Japan

[covert wolf]

I haven't been able to create Sora 2 videos for a few days now, what's going on? I'm so frustrated!

[coral aspen]

If this is a plot to force Plus users to switch to Pro, I think it is stupid along with the person who came up with it

Why would we not just switch to a different provider after being betrayed like this?

[coral quest]

Hello, I’ve had a problem with ChatGPT for two days: when I send it images (screenshots, photos, etc.), it tells me that it sees nothing except a block of text made only of repetitive characters, like:
_àâbc_âabc_âabc_âabc_âabc_âabc_âabc_âabc_âabc...
(a very long sequence of the same pattern repeated hundreds of times).
Yet my image is normal, and it worked a few weeks ago.
Do you know what this could be due to?

[glad coral]

Having this issue for two days know on one of my chatgpt accounts which it doesn't generate or respond.

[gloomy pumice]

Whenever I try to use a 1 month free deal for pro, it declines my card, even with the correct amount of money on the card. How do I fix this?

[cobalt anvil]

How do I report an outage when Sora can't generate? The status page says everything is fine, but it's not

[wooden meadow]

Help 😭

How do I log him out

[trail cairn]

https://help.openai.com/en/articles/7242626

[south solar]

How long is "Identity in review" after verification is successful on the phone?

[drifting cosmos]

Do y’all happen to know, if I make a custom bot, can non-paid accounts access it?

I’d like to make a bot that my students can use to get more practice on what I’m teaching them, but I don’t want to make them have to pay.

[low orbit]

Hi everyone, I have a question about Codex, so if I use Codex will it impact into my chatgpt personalization memory profile which I have in my chatgpt?

[supple hatch]

Nope

[trim jackal]

Is chat having issues lately??

I keep getting stuff like this when I upload images. Or if I'm asking it to do something it will tell me the request was flagged as sa scam or fraud..

Every image gives me the same answer

Tried multiple different images. And it started entirely new chats and it doesn't correct the issue

[tranquil socket]

Does anyone recommend courses for getting into ai? Is coursera a good option? Open to other platforms 🙏

[trim jackal]

How do I fix it?

[thin egret]

does anyone know why this is happening when I try to make a character in Sora ai

[vivid mortar]

Depends what you are looking to learn! I really dont think something like coursera is necessary, just youtube alone will have plenty of content and likely more up-to-date information :)

[lavish hound]

how can i use sora 2 from europe?

vpn to usa?

and then what?

i still cant

nvm

i got it

[coral quest]

I have the same issue

[stoic wasp]

Need to verify your number now.

[lavish hound]

im good ty

[short jolt]

Not sure where to ask this, but we're trying to use OpenAI for a project, but need to access the data residency options, which, according to the FAQ, use the contact sales form. However, whenever I submit that form, I just get a generic "use the self-help services" and can't get in touch with anyone. Not sure if there's anything else I can do? Has anyone else been able to get this feature enabled?

[sleek garnet]

|| e ||

[tawny hawk]

whot

what

[next tusk]

hey, I really need help.

my company is shutting down our ChatGPT Business workspace and I’m about to lose access. I’m trying to separate my personal ChatGPT account from the Business workspace before it’s deactivated, so I don’t lose my own work.

right now everything is tied to the Business space, and support told me there’s no export and no recovery once access is gone.

is there any way to detach a personal account from a Business workspace (or convert it back to a personal account) before shutdown?
I just want my personal data to survive after the Business workspace disappears.

any help from someone at OpenAI would seriously mean a lot.

[trim jackal]

Why pay for a subscription if this tdosnt even work...

Like still an issue. 0 response about it. And it's been weeks.

[timber shadow]

Hello, according to this article, it is not currently possible to reverse a migration/merge of business/personal workspaces:
https://help.openai.com/en/articles/8801890-can-i-migrate-or-merge-my-personal-chatgpt-workspace-over-to-my-chatgpt-business-workspace

Please note that merging your workspaces is irreversible.
If you migrate your data, that data is now considered part of your Business workspace. If you lose access to your Business workspace, you will also lose access to the migrated data. This means that if you ever leave or get removed from the Business workspace, or if the workspace itself gets deactivated, you will lose access to all of your data in your Business workspace.
So you will get a new personal workspace following the deactivation, but it doesn't sound like de-migrating/de-merging is currently possible.

[timber shadow]

Hi, have you tried attaching an image on another device or platform, like on chatgpt.com instead of in the app, or on a computer?

[trim jackal]

Tried on the app, Google, Firefox, and on both PC and mobile.

I've also been able to recreate the exact same problem in every single chat I've tried it in.

As I said this has been an issue for weeks and none of the "fixes" have worked

[timber shadow]

Very odd! Have you also already tried reaching out to OpenAI support about this? They may be able to give more info. It's the chat button on help.openai.com, in case you haven't tried yet.

[trim jackal]

Tried using the report function for the chats and the only option it gives me are things like "something illegal, something against minors" and what not.

That doesn't seem to be any form of feature for reporting errors 😅

[timber shadow]

Ahh gotcha, that makes sense. The official channel for OpenAI support is the chat button on help.openai.com, more info here: https://help.openai.com/en/articles/6614161-how-can-i-contact-support. You will likely start by chatting with a bot about the issue, but in my experience, a support ticket will be created for human review if the initial troubleshooting steps do not work

[tranquil socket]

Hi! Thank you so much for the response! Im looking to learn how to generate applications I have done quite a bit of course work within Coursera but started to get lost when SQL and Python were introduced. I know the basics of ChatGPT and Vector Search. I have been able to produce an application but unsure of next steps from there.

[vivid mortar]

Look into cursor and Claude code or OpenAI codex

These tools will help you massively when building complex projects with multiple files so you don't have to copy and paste from chatGPT

[tranquil socket]

Thank you so much!

[grizzled pagoda]

Hi Guys, what is best alternative to 4.1 mini. I have been using it for a long time in my voicebots. But lately the performance degraded a lot and it feels like its very dumb now. Any one want to suggest any latest model in the same cost range. Would like your experiences with other models.

[summer ruin]

hey guys, i need some help with my year in review, anytime it gets to 100% nothing happens, and the whole time as its loading the "retry" button is just sitting there, when i click "retry" the loading bar starts from 0 again and the retry button comes back then the widget loads till 100 again and does nothing...

[rigid moat]

Hey! Quick question about Audio Native.

Is there a way to exclude specific HTML elements from being parsed? For example, we have image captions like <figcaption>Photo: John Smith</figcaption> that we don't want included in the audio.

I've checked the docs but couldn't find anything about exclude selectors or skip attributes. Is this possible?

[trail cairn]

wrong discord

[coral aspen]

I'm literally just trying to generate pictures for a class I'm going to teach and I'm getting rate limited

On pro

Im cancelling my subscription next month and buying one for G****i (didn't mention a competitor here and you cant prove otherwise ☺️ )

Maybe Ill even ask my bank to charge this garbage company back because I deserve free prices for the free tier equivalent

[thin kite]

What happened to the Codex channel??

[raw pewter]

Is there no way to add project limits??

how do I modify the limits field

https://help.openai.com/en/articles/9186755-managing-projects-in-the-api-platform

looks like you can only do soft limits, smh

[hollow hill]

I am trying to receive chatgpt notifications on my phone, for some reason I can’t get them. Can anyone help me with my problem?

[green summit]

GPT app > settings > notifications.

And in app info (android, not sure about IOS but I assume it's the same-ish)

[stable grove]

Will this offer ever be available again?

"One year of ChatGPT Plus
free for transitioning
servicemembers"

[glossy nest]

Does anyone know why are my available credits and my budget so different ? I only put 5$ once recently, and used my key in a python script, the january budget is my actual usage of the api, however information on my billing page are not matching

[rustic marlin]

budget is the value you can manually set on the dashboard with 2 limits
the lower limit where it sends you a email warning the suage crossed it, and the upper limit where it blocks the suage when it reaches it

the budget is not connected with the amount of credits you have purchased

the budget is just a safe guard to prevent over-spending

[glossy nest]

Yeah i get that, but why my budget was getting up when I used my api key and my "credit grants" were remaining unchanged ?

like rn it means everything I did in my python script was for free

the 0.13$ i see on the billing page is actually when i did some tests on the openai dashboard, it's like only this was debited from my credits and not my usage with python

i don't know if i'm being clear haha sorry but it's so weird