#community-help
1 messages · Page 199 of 1
even in private mode that still has't worked
i don't get it. it worked a few days ago. and now they don't work
So if anyone can help. Thank you.
I still see them at data controls/view archived chats
Also I think they are only there for 30 days or something before deletion (if I recall correctly)
No they have been there for a long time and they are not deleted.
They were openable days ago and now they are not.
I was able to view what is in the chats days ago and now they can't be open.
This has been happening to most of the archived chats except a very few.
Anybody that can help me?
Did you try logging in through an incognito window (or another browser) and try again?
Yes I used Firefox incognito private browser and that still didn't work.
This is what I'm getting.
I hope anyone else can respond. I need help.
I hope no one ignores me.
I need help
Please do not ignore me.
hello??
hello?
my codex has been thinking for 6 hours and 30 minutes should i stop it
its just having some shower thoughts
Yo,I’m having to write a mediation which I have done with the help of ai now. Does anyone know any good humanizer or could humanize the text for me ?
Anyone know why when i open a new tab of https://chatgpt.com/, the text box gets automatically filled with the last request i gave it?
it for sure already timeouted long ago, redresh your browser
You should definitely reach out to OpenAI support
1st image: that cannot be fixed, unless you edit a message that was sent earlier, but that will still go until the limit is reached. So you have to start a new chat to continue, which sucks, I know.
2nd image: yeah, clearing cache and site/app data might help for that.
Data on browser:
Open browser > settings > find "site settings" > then "all sites" > search for "OpenAi" and "GPT"> delete data > then reopen the browser and try again
Data on App:
Go to the homepage on your phone > hold your finger on the GPT app icon > select "App info" > find storage > "clear data" > reopen the app and try again.
(Don't worry, none of this has an effect on your account data)
(Cache can be found next to Data in app info, and you'll find cache nesr site settings, too)
If all that didn't help, contact support.
And close your damn tabs!
The number...
What tabs just so I'm not confused.
Seriously what are the tabs?
I suspect Amnesia meant 'browser tabs', but I'm not sure.
Correct. ^
@glad coral
Hi, I'm trying to figure out how to use Openai api, to generate images with GPT-image-1.5. It seems like it is possible that attach reference images as input for client.responses.create(), but is it possible to do that with client.images.generate()?
Hey everyone!
Lead Pulse is LIVE on Product Hunt — local B2B lead gen from Google Maps + reviewed outbound.
Free Starter plan available!
What do you think of the review-first workflow?
Hey everyone, is it just me or is ChatGPT super broken lately? I’ve been running into a bunch of weird issues and wanted to see if anyone has a fix.
First off, my entire sidebar with all my chat history has just vanished. To make things worse, the Web and App versions are totally out of sync—they’re showing completely different things.
Also, every time I try to start a new chat and send a message, it just glitches and kicks me back to a blank "New Chat" start page. It’s like an infinite loop.
On top of that, whenever I try to open an old chat from about a year ago, I get hit with that "The conversation is too long, please start a new one" error.
Is this a known server issue, or is my account just cursed? Any help would be appreciated!
like this:
I've got around 200 chats in my history, but right now I can only see a handful of them. The rest just won't load at all. I’ve already tried logging out and back in, and even reinstalled the app, but it only made things worse. I rely on these conversations for my work and I’m really stuck here. 😭🙏 Has anyone else run into this? Any help would be huge!
this has always been there for years, for all plans. when you reach the limit you need to start new one, also long chats can hang and slow the page
I avoid those problems from affecting new chats by frequently creating new ones, I carry over info I want into the new. Chats start to act weird, usually, before they reach max length; if I notice slow downs or other oddness, I take that as a strong warning, time to stop that chat, so I can still access it and move what I want to a new chat.
Hello everyone, I just created my ChatGPT business account, but its acting wierd, it gave me an error while creating the workspace (did get charged and my account does say business) but the console is bugged, wont let me change anything its just blank, wont even allow me to add new members
Do I need to wait or should it be working right away?
how to make chatgpt website follow custom instructions
i have
- use very little emojis
and it used none
its very inconsistent
chatgpt in the past years ago followed instructions much better than it does now
what happened?
Enter custom instructions > use the thinking model so it'll remind itself to follow its instructions.
That's been the most consistent way for me
custom instructions - personalization section of the profile
Oh right, I forgot to add that. thanks!
I'd reach out to OpenAI tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support?
Random query so I see you can make a privacy request for all your data which is apparently processed and you get sent a download link of a zip. Does anyone know if this request means a human has to review all the data thus removing any possible security/privacy, or if it's automated.
it's automated with human oversight, sometimes an OpenAI person may review the request but it's usually automatic
Is chatgpt down or something? Cause my app isn't loading
working fine for me
Guys my subscription dissapeared
I bought it week ago
Does anyone have same issue? it says im on free plan.
No
Why doesn't it let me use any other model when I have pro on website lmao
I hit my usage limit. Codex says wait until April 19 OR upgrade to Plus which I did, still says I have hit the limit. What is the point of upgrading? Am I missing a trick?
I wonder if it's because that same chat in that same project has done 'something' that requires 5.4 to see/work with. We used to have some chats that would lock into a specific model after certain tasks were done with them, this looks kind of similar.
If you're interested, check a different chat or create a new chat in the same project, if those can use other models, chances are yours has used some tool that requires more of only 5.4
Opened a new chat, it still telling me to upgrade to Plus, I already have upgraded to Plus.
I don't use Codex, I'm not sure about your question. You can reach out to OpenAI tech support (we're community members here, nearly all of us, only the gold named work for OpenAI, #server-staff ) through https://help.openai.com/en/articles/6614161-how-can-i-contact-support?. I hope they can help you!
Need assistance? Here's how to reach us and get the help you need.
I've tried to do this in a new chat and same project, but it's locking itself to this only.
Tried on my Macbook and desktop too
desktop and macos app works fine
What do you see when you click that drop down arrow?
Also, what do you see when you click this drop down arrow?
This is a brand new account btw, I couldn't upgrade my plus subscription in time as it used apple pay and I used a different card not linked to my apple account so i made a new account and paid lmao
Works in both apps though properly
I've turned off ad blockers, tried raw browsers like Chrome and Safari, different devices, seems to be a exclusive web browser issue?
Hi, I'm trying to sign up for the Cyber program at chatgpt.com/cyber. My verification was deemed successful by the external verification platform, but upon coming back to the ChatGPT page, I get this message: "Your identity couldn't be verified or your account is ineligible at this time. If you think this is a mistake, please contact support." Anyone know why? I also can't seem to contact support, the chat bubble doesn't show up on the help page. Thanks for any help!
can anyone tell me why codex is so damn slow for me
just buyed the pro plan for 100 usd generated 1 picture and now it tells me i have to wait 10mins i hit the limit
did i missed something? Oo
The app is brutal slow today 😩
what help do u need?
I would restart my computer, the no chat bubble showing up bottom right sounds maybe like the sort of issue that might help fix. And/or check for browser extensions, might need to turn something off, possibly.
The verify issue maybe is cache related (note, I'm a volunteer community member, offering ideas I have, I'm not tech support, not even that computer skilled, just have heard others talk about problems and fixes rather often), otherwise if a little time doesn't fix it (maybe system needs to update and hasn't had the chance yet) I'd definitely contact support.
Just to ensure you have the correct page for OpenAI tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support?
Need assistance? Here's how to reach us and get the help you need.
interesting
Why is the ChatGPT Windows App so slow when the chat is already long, but on my iPhone e.g. it's fast and not laggy/buggy at all?
trying to upgrade to pro i have the money in my account but it keeps coming up with an error "There was a problem updating your subscription"
@timber shadow
Hey! For account/billing assistance, I'd recommend reaching out to the OpenAI team for support. The best way to reach out is the chat button in the bottom-right of help.openai.com. You'll start by talking to a bot, but the bot can escalate to a human for review & assistance if needed. Hope this is resolved for you soon!
is anyone else’s button not working?
Once verified, how do you get access to 5.4 cyber? I've been verified since feb.
Hmmm something didn't look right with your request. Please try again later or visit ¿What is this problem? help me pls
Can soemone explain to me why I cannot see 5.3 Spark on my Pro 5x ($100) account
Do we not have access?
I thought it was on the pro tier
or only $200 plan
Despite me having Plus, i can’t seem to use any other models than Thinking Mini (tested both on web and in the iOS App), anyone knows why? 🥲
Why is the ChatGPT Windows App so slow when the chat is already long, but on my iPhone e.g. it's fast and not laggy/buggy at all?
Anyone with a fix for this?
easy fix is move to macOS
I noticed something peculiar yesterday. After not having used the Codex app or CLI for a while, when I opened the app and started using it, a bunch of models were missing from the selector. I checked the CLI and found the same thing. Then I checked online, and some people suggested logging out and logging back in. So, when I went to log out, in the bottom left corner of the Codex app when you click the settings button, it shows a little panel showing other buttons and login info. It said I was logged in using an API.
I logged out and logged back in. My email then showed in the settings pop-up menu, indicating I was now logged into my personal account (the correct login), and the missing models reappeared. I did the same for the CLI.
Something they did must have bunged up the authentication for the app and CLI, leading people to end up using the API instead. Maybe this could explain the usage issue people seeing, as people could be logged in using API and not using the ChatGPT subscription.
Fair advice.
Source: ex-Windows user since 2008-ish
Hey I need help, my previous OpenAI account got locked and I can't transfer my payment stuff to my new account in order to upgrade the subscription, how do I go about fixing that?
I had this same question but on my Mac.. it’s so freaking slow but my iPhone is fast
Is anyone else experiencing massive looping behaviour from GPT? As in, referring back to old unrelated input/prompting and not the recent one, over and over in the chat? I'm using 5.4 Thinking, and this behaviour just suddenly started yesterday or so and makes the AI near to unusable.
What status do you need to be accepted into 5.4 Cyber and be able to use it to its full potential?
Ea$y fix yeah 💰
Idk why the app is so different than our phones one
hello i have a problem, i am writing a exam tommorow and just bought chatgpt plus however the chat is not in plus my account is but i cant write anymore or send fotos
pls help me
when i try to upload a screenshot from my clipboard it says unknown error and wont let m,e
mid2015 mbp with 16gb ram && AMD gfx going about 199eur in baltic EU
2021 mbp with m1 pro/max && 32g ram going about sub-1100eur in baltic EU
yes, this is totally "rich man's" spend
So, I have codex because I have GPT plus, since you guys released that new plan, maybe im imagining it but it feels like the limits run out alot faster, is the new plan does that consider what I already pay for Chat GPT plus or no?
When will GPT 5.4 be available in Indonesia?
It already is, you just have to be a Plus subscriber or above.
GPT 5.4 Thinking is only available with subscription.
guys how do I join the announcement channel
it says you don't have webhook to join server
Not available in Go?
I looked it up,
Go users get access to 5.4 Thinking Mini (but with less messages obviously), Plus and above gets access to the main 5.4 thinking model, and Pro users also get access to 5.4 Pro.
So yes-ish
What's OpenAI's support contact? I bought some codex credits a few days ago, recieved an email today that I was getting extra credits due to some platform changes & now I have 0 credits in my balance lol.
Guys where can help me
I’m using the Codex app for coding.
I have a chat, and when I click “Fork into new workspace” in my current chat, it creates a new workspace with the same conversation tree as before.
But when I go back to the original chat, I can’t find the workspace I just created. I also checked my chat history, but it’s not there.
Hi everyone, I have a question.
The Study and Learn feature disappeared from ChatGPT. I used to have it before, but now I cannot see it anymore.
Has this happened to anyone else? Do you know how to fix it or get it back?
Thank you.
happened to me too
Hello,
I am contacting you regarding a serious and unexpected billing issue on my account.
I recently added $20 in credits, and my balance was over $10 with no outstanding charges. However, shortly after that, I made only a few API requests using GPT-5.4 (approximately five small requests).
Within approximately 3 minutes, my usage suddenly increased by around $55, bringing the total to $77 and resulting in a negative balance of $47.54.
This behavior is highly unusual and does not align with my actual usage. The requests I made were minimal and should not have cost more than a few cents (less than $0.01 each).
I am concerned that there may have been:
Unexpected or abnormal API activity
A potential issue with usage calculation
Or unintended usage beyond my control
I kindly request that you:
Investigate my account activity during this specific time window
Provide a detailed explanation of how these charges were calculated
Confirm whether there was any abnormal or unauthorized usage
Consider adjusting or refunding the charges if an error or misuse is identified
This sudden increase in cost within minutes is very concerning, and I would appreciate your urgent assistance in resolving this matter.
Thank you for your time.
nobody here can help with this, you need to contact support at https://help.openai.com
I need an email address so I can contact them.
you have to go via the support site
I don't know how to use the website. What should I do to get my balance back to what it was?
the support site is where you will get an answer to this, we can't help you here
I don't know how to use the website.
This is a problem, but to make it easy for you: https://help.openai.com/en/collections/3943089-account-login-and-billing
Hi everyone. I can't save any new memories right now, is there any known problem with this today?
I've saved two memories in the past hour
hmm..Strange.
guys if i have been using ssh codex with my business account how can i switch to api without logging out?
What is happening to in chat branching? All of the story variants I made are gone!!
Bring then back!!
Does codex cli have a yolo mode like qwen?
yes it does codex --yolo "your task here"
this is short for codex --dangerously-bypass-approvals-and-sandbox "your task"
the param should be named --dangerously-bypass-approvals-and-sandbox-and-i-will-not-complain-on-discord-when-things-go-bad

Hi, I found that the Prism OpenAI service have been discontinued, but I was not aware of this change in advance and did not back up the data I had stored on the platform. Where can I get help?
https://help.openai.com/en bottom right
thanks
I am on a 2gb ram/50gb disk DO VPS. Do you think it is enough(I don't) to support VScode AI extensions? The connection drops often...
hi guys How to download the whole thing from Sora 1?
hola! Is there a channel for chatGPT app feedback? like for feature requests?
perfect! thank you @onyx kraken
Very sus
they don't even try atp🙏
vscode ai extensions shouldn't be very heavy?
are you running the vsc server on it or what
goated email domain though
maybe check out hetzner shared compute or github code spaces
that machine would be struggling
For the past few days I've had this problem. I cannot upload a single image to my ChatGPT iOS app. Can someone help me please?
this is bad, i was excited for prism
this is indeed not an email from OpenAI
kinda.. codex runs on node.. a running node instance might eat 200~300 MB of memory.. which is probably too much for what it does but works nonetheless
Is anyone else having login issues?
Yeah Its down
NOOO
Same here... Portugal but I get Cloudflare error: Cloudflare Location: Madrid
Error reference number: 504
Well im going out, hope its fixed by the time I'm back
We are investigating the issue for the listed services.
Welp
With sora beeing shutdown will chatgpt get sora capabilities for generating images? I made the same prompt on chatgtp and on sora 1 and sora results are way better.
I switched to a business account and now I can't log in. Plus, I didn't realize that it would creat a second workspace. That wasn't clearly defined, and now I can't even reach an agent! Ugh! What do I do?
Hello, when someone tries to overwrite another message, the following message appears: "Editing will permanently overwrite all messages after this point"?
We are continuing to investigate the issue for the listed services.
My issue too! How do we get the dang arrows back?
It works fine on my alt account
I want them back DANG IT! Get rid of this permanent stuff
Yes, I would like him to come back too 😕
I mean if it was universal, I'd understand it's all accounts but no. My alt account which I started like a month and a half ago still has the numbers and arrows!!
Like why my main ?! I had so much on there!!!😭😭😭
Just a quick question, for ChatGPT Plus, I read that thinking usage has a usage limit of 3000 messages. Does it matter if I'm using Extended or Standard thinking? Does Extended consume more usage?
hello every_one, I started to use tools.web_search ? and I'm trying to understand the things behind this tool
Like what realy happen when the user enable the web_search tools ?
Can anyone help me with an overview of all AI and what i need to know
I had GPT make an overview for you!
Here’s a simple overview of AI and the main things worth knowing:
1. What AI is
AI is a broad term for software that can do tasks that normally need human intelligence, like writing, analyzing, recognizing images, generating music, coding, translating, and answering questions.
2. The main types of AI people use right now
A. Text AI
Used for writing, summarizing, brainstorming, explaining, translating, coding, and research help.
Examples: ChatGPT, Claude, Gemini
B. Image AI
Used for making or editing images, art, logos, thumbnails, concept designs, etc.
Examples: Midjourney, DALL·E, Flux, Stable Diffusion
C. Video AI
Used for generating or editing videos from prompts or images.
Examples: Sora, Runway, Kling, Pika, Seedance, Dreamina
D. Audio AI
Used for music, voice generation, voice cloning, sound cleanup, transcription
Examples: Suno, Udio, ElevenLabs, Whisper
E. Coding AI
Used for helping with programming, debugging, explaining code, building apps
Examples: ChatGPT, Claude, GitHub Copilot, Cursor
3. Important words to know
Model: the actual AI system doing the work
Prompt: the instruction you give it
Context window: how much info it can remember in one chat/request
Training data: the data it learned from
Fine-tuning: extra training for a specific job
Inference: when the AI gives an answer
Multimodal: AI that can handle text, images, audio, video, etc.
4. What AI is actually good at
Explaining things
Summarizing long content
Brainstorming ideas
Writing first drafts
Coding help
Translation
Image and video generation
Pattern recognition
Speeding up boring work
5. What AI is bad at
Being consistently correct
Understanding truth from confidence
Handling vague prompts well
Giving reliable facts without checking
Deep human judgment
Complex reasoning without mistakes
Knowing current info unless it can browse/search
6. Biggest thing to understand
AI can sound confident while being wrong.
So: use it as a powerful assistant, not as an automatic truth machine.
7. Best way to use AI
Be specific
Give context
Say what format you want
Ask it to improve or rewrite
Double check facts if accuracy matters
Use follow-up prompts instead of starting over
Example:
Bad prompt: 'make this better'
Better prompt: 'rewrite this in a clearer, more professional tone, keep it under 150 words'
8. The main risks
Hallucinations, meaning made-up facts
Bias from training data
Privacy issues if you paste sensitive info
Copyright/legal gray areas in some cases
Overreliance, especially for school/work
Fake images, fake audio, deepfakes
9. What beginners should focus on
Learn how to prompt well
Learn which AI is best for which task
Always verify important info
Do not upload private/sensitive stuff blindly
Treat AI output like a draft that still needs your brain
10. Easiest way to think about it
AI is less like a magic brain and more like a super fast assistant.
Very useful, very powerful, sometimes brilliant, sometimes confidently stupid.
11. If you only remember 5 things
AI is a tool, not a person
Different AIs are good at different things
Better prompts give better results
Always fact check important output
Human judgment still matters most
Oh damn.m
Might have to send you in dms.. this is absurd long
Sent
please i also have other questions
Sure, what's up?
Please help, script can't read anim_processed v2, it curses skeleton and animation is broken
a bit of explanation - visual is skeleton like armature, mesh is primitives, I can send it if anyone wants, and fbx is animation og format which is not always available
entire animation is just man sitting and nodding slightly
why is it im getting security errors generating images of more traditional people or anything trademarked? super frustrating
defitnitly not the same generator as a few hours ago
super frustrating it shows how much it is going to blow and be weaken over time
translated message: “Sorry, but the request may violate the safeguards we have in place regarding similarity to third-party content. If you believe this is a mistake, you can try again or edit your request.”
thank u for nothing
Does somone know why im stuck in this screen
i still need help can someone help me? why chatgpt creating pics for me when im saying i want a text prompt… i have this issue lately not creating texts… just creating a pic instead.. have this issue for days now
It's a bit buggy rn with the newest update. I've had it happen a few times, too. But I usually cancel and tell to strictly use text. Which is also a hit or miss atm.
do u think that will be fixed? :D
Yes. It will most likely be fixed soon. (I wish I could tell you when, but I'm not sure, but what I am sure of is that they're aware of it :)
that buggy so pmo xdd
atm
🥲
this bug came when the new layout came right so few days ago? 😔
Yeah, that was yesterday, but might have become a bug a few hours before the update. Could possibly be tomorrow or so.
Are you web or app user?
app :p
Alright, yeah same (for the most part). Most likely an update in store soon.
Or maybe this is going to be one of those things it takes years to fix 🥲.. I mean, I'm kinda still waiting the GIF issue gets fixed since last year August..
On mobile, you cannot send a GIF, when trying to upload it'll fail so you cannot send it, but when on PC or using the mobile web version it'll just upload.. all of sudden it was broken and never looked at again by OAI.. meanwhile, a gazillion mobile app updates were released..
So yeah, but I'm sure this will be fixed soon because this is more of a priority rn because of 2.0.
👍
Anyway whistles
[urgend]
I need confirmation from OpenAI staff or anyone with verified information.
I am trying to preserve my old Sora 1 assets before the shutdown. I requested the official export multiple times, but the archive only contains a JSON file and one Sora 2 MP4. My older Sora 1 media are still visible in the Sora library, but they are not included in the export.
I have also submitted a Privacy Request, but I do not know whether it will actually contain all Sora 1 files.
My question is:
Is there any official bulk download/export path for Sora 1 assets, other than manually downloading every file one by one?
More specifically:
- Will the Privacy Request include all Sora 1 media files?
- If neither does, what is the officially recommended way to preserve Sora 1 assets before the deadline?
I am not looking for guesses. I need confirmed information because this is a data preservation issue and time is limited.
-- Manually downloading my entire library would be a nightmare, since the download process is very slow and sluggish. Selecting multiple files isn't handled properly either. I've tried that before, but it's a joke—nothing works (I'm still using the "old" Sora 1 web interface since I'm in Germany).
Yeah it's a bit confusing right now. Sora.chatgpt.com/exports/me is the official way to "preserve your content". However, it seems to be only for Sora 2, and not Sora 1. Which is confusing because it should be both..
There might be some tools made by the community that can help you get all images/clip from old Sora, I know there are a few for Sora 2, but there's most likely also one for Sora 1.
my voice mode is not working by some days in my ios app. i tried disabling the advanced voice mode, i tried changing the voice, nothing works, really needing help
i even uninstalled the app and removed the data
Maybe a dumb question, but could possibly your mic be disabled? Have you tried a different device?
i even re-enabled it(the microphone)
Hm
and after reinstalling it asked again to enable, i clicked to accept it
Are you on PC or Mobile?
i have no other ios device lol
Mobile, on IOS
Check your phone settings, it could be that your mic is disabled. (Like. I checked off both my camera and mic, until I need them, so maybe you accidentally pressed it or something)
i have no button of microphone on my iphone
😅
Thanks for Your reply @green summit . Unfortunately, that doesn’t really help me either, because the deadline is coming up very soon and I still have a lot of data to back up before the service is shut down. Sora’s download feature is really terrible—it takes forever just to load the gallery. An all-in-one exporter would be really helpful. Actually, OpenAI should also take care of the old Sora 1 data so that users can act in time without getting stressed out (due to long loading times and the lack of an export function, which has been requested for ages and still hasn’t been implemented)
no offense, just feedback!
I'm not really a fan of AI generated message but alright I'll respond.
Well, that's why these tools might help you out, I'm going to try to find one for you, that way you can download it all in one go.
But yeah, it's a pain to download it all one at a time.
I've already contacted support (https://help.openai.com/), but no one has responded and no one can help me.
https://www. reddit .com/r/SoraAi/s/VF15EejSln
[SoraVault] is one made by the community due to Sora going away.
Yeah, support won't be much help anyway
This is probably your best bet.
Anyway, I hope that's helpful and you'll be able to get it all before the end!
I'll give it a try. I got your link. Thanks for that. Unfortunately, it doesn't really work for me, but I just checked Reddit and searched the subreddit for "mass downloader." There's also mention of a Chrome extension there. I'll test it all out; maybe we'll get a response from support eventually. Otherwise, it'll be a bit of a hassle. Thanks anyway for your time and your “Let me Google that for you” approach 😉 It’s kind of obvious, but right now my heart’s racing from the shock… so my brain doesn’t seem to be working properly. Thanks a million! Have a great week
***Your recommendation was worth its weight in gold!!!
Nah you're good! It happens, things can easily be overlooked really.
And appreciate that, you as well, my guy 🤙
On Android, they just released an update, might wanna check on IOS too and see if it still does it for you.
Even though it's for "Agent Workspace", it may also include bug fixes.
Yo does anyone know if theres a local voice traîner for singing for mac without gpu, ive spent so long and cant seem to find one✌️
no update on ios yet
nvm was bugfixes update earlier today so maybe have been fixed
Any fix for this billing issue: https://community.openai.com/t/unable-to-upgrade-to-chatgpt-plus-payment-authentication-error/1378796
Hi Team, I am trying to upgrade my ChatGPT account to the Plus plan, Previously used Free Go Plan, but I am consistently facing a payment error. Error Message: “Your card may be invalid or authentication may be needed.” Details: Country: India Payment Method: (mention your card type – Visa - Debit) Platform: (Web / App) A...
nope still buggy @green summit , the bug still there when i just want a text...
I've been working with multiple vscode codex threads (of different workspaces) at the same time in the same machine for the past month+,
I've noticed this was very costly so I upgraded my subscription to pro recently.
starting this week specifically this started causing hangs on all those threads when I try to use one of them, then the other too while the first one is still thinking - opening a third is not even possible, it just hangs trying to load that thread.
anybody else suddenly stubmled upon this behavior change? is this tracked and recognized somewhere? is there a fix upcoming? anybody knows of a good workaround other than migrating to a compettitor?
hi need help anyone know good ai courses for beg?
I am sure his has been discussed - is there a reason sora / chatgpt exports are never arriving? I asked for one nearly a month ago, and then another two weeks ago.
Clips: when you go to sora.chatgpt.com/exports/me (make sure it says "exports" instead of "export") > click "export" and wait for a little bit > then after you get a mail go back to the Sora.chatgpt.com/exports/me [manually. The link in the mail doesn't always work] and you'll get a list with zip files that you can download. This will contain both posted clips and drafts. (Posted clips will be in the spot where it'd be if it were still in drafts, it'll be all in order)
This is different from the "Export Data" button in your settings. Which will only give you GPT related stuff.
Thankyou. I know how it works but I never get the email telling me the export is ready
Yeah I heard some people having some issues with that.
Try going to the link (sora.chatgpt.com/exports/me) manually, if you clicked "Export" on that page before, it might now show a list of zip files.
Great I will try that. Thankyou
Hello, does anyone know how to get listed or rank among sellers for a product?
Do you need to provide a specific schema or structured data, or is this pulled from paid results (e.g., Bing/Google)? Is there a good way to provide all these price, shipping, return policy data to ChatGPT?
I’m aware there’s a way to upload a product feed, but as far as I know, this option is currently available only in the US (I’m based in Poland).
Someone know why I can't select gpt-image-2 in this interface ? Yesterday It's was working fine
anyone please?
restart it, but end process in task manager first
It's on web
could be cache causing issues, dunno
Is there any way I can revert from image v2 back to v1 or whatever the previous one was?
I don't think so, however, if you really need it, the older model still on the API
in fact, you can use it on #image-bot
Im yet to update this bot in the next few days to add the image gen v2 xD
Thanks a lot ill have to try that
i cannot use the api whatsoever with my subscription right?
thats why im asking
no, chatgpt subscription does not include API usage, you gotta pay for that
ye alr
problem is ill have to setup some stuff myself to use the api, cant just go into the chat
not sure how much time thatll eat
well, unless you really need a high volume of images in the old model... you can get a fairly large amount of uses jsut by loading 5.. 10 bucks on the API
yea thats not an issue if i can get the results i need
ease of use matters tho
it is not cheap.. but if you just have a situation where you want it for some reason.. you can get some without spending a bank breaking amount of money
image v2 broke transparent image generation
rn the best workaround i have is generating with greenscreen and removing said greenscreen
after all, in the API, you fay for exactly what you use.. so.. if you need to use 5$ worth of AI.. you pay jsut for that.. you don't need the whole 10$ subscription
but its still not as good as it used to be
hmm, interesting, I have not tried trnasparent images...
I'll have to check that soon when I update the @gleaming herald =P
because it has a setting for forcing the AI to do transparent images
i spent over 5h testing already, only way to make it work is selecting INSTANT model but that makes the quality suffer
I'll have to see how broken it is..
just look at him : )
btw.. are you sure "transparent images are broken"? what is your prompt?
yes
i tried a lot
i tried mayn different wordings
true alpha channel
subject cutout
export as transparent png
remove background
etc
all works fine on instant, but thinking gives you the fake checkerboard thing
you know when youre in photoshop or wtb and theres no bg, that checkerboard you get? it just generates that
hmmmm, that does not seem right
yep
its not just me ofc
other ppl on reddit posted abt it
search in this server for "transparent" some ppl alr talked abt it
here.. the PREVIEW has the checkerboard background, it is not on the image if you download it
did you turn on thinking
it was on thinking in light effort
this prompt worked fine for me
maybe its only on more complex prompts where the issue appears for some reason
gimme prompt
the prompt is what you see on the screenshot
memories are disabled
ik i just tried it
send me the prompt you used and also the model and settings
yea if it doesnt work now ill send it
you dan DM me if you rather not share it to all
got it
try that
@rustic marlin
apaprently, v2 does not support transparency yet..
https://developers.openai.com/api/docs/guides/tools-image-generation
I assumed it did.. but well...
¯_(ツ)_/¯
the AI probably used v1.5 here
@ashen mantle well, there are many services for extracting the background, some are AI, some are not even AI, you might try searching for those
Yea I've been trying a few, best option seems to be green screen
so weird
getting these failures recently at increadibly high reproduction rates, started this week on codex vscode. any way to get around this? any ways to prevent it and stop it from reproducing? staring at my chat clicking send waiting for it to end differently like a russian roulette without empty bullet sockets firing every time
how to know which model chatgpt image is using?
hmm wdym
how do you know which model you're using for image creation? like I saw some ppl dont know if theyre creating images on the latest model or older. tbh im not too sure either
The new model is the default.
@green summit im still getting pics when i just want text prompts 😬😬
🥲
My app keeps telling me Ive run out of free chat with image generation but Im not making or sending photos and it says to make a new chat but then just keeps saying the same thing when I do. Help please @ me
uhhhhhhhhh
I submitted an appeal.
Am I not allowed to make a new account? I'm now banned from ChatGPT I guess?
Oh Noo That So Sunds Awful to Hear
I'm really confused why this happened
Do you have Any Inkling that could have cause this
no..
also this server requires an account linked, so does that mean i'll be banned from this server since my openai account was suspended?
Maybe i Dont Know
Hi, we can't advise on anything related to your OpenAI account - you can reach out to the OpenAI team via the chat button help.openai.com (or by appealing, as you noted). I'll be removing some messages about this per Rule 2 of the #server-rules "All content must be suitable for all ages."
For any questions you have about server moderation related to your account, send a message to @burnt galleon and the mod team can review & respond 
Hey all! I am a small business owner and I use Codex everyday and am using the $200 plan. Last week my OpenAI account was suspended for no apparent reason, I think maybe I was using it too much? But I am in desperate need and OpenAI has not responded to my appeal after an entire week! Does anybody have some reccomendations on what to do here? I am literally begging for help at this point. If anyone from OpenAI sees this, please help! I use Codex for everything and need it!
What's the suspension reason?
My account was suspended too... Mine for...
they didn't give a reason yet
What did the email say?
this: Hello,
OpenAI's terms and policies restrict the use of our services in a number of areas. We have identified activity in ChatGPT that is not permitted under our policies.
We are deactivating your access to our services immediately for the account associated with the email.......
This is super disheartening considering I use Codex for everything. Also I spent several hundred dollars on this account
where you uploading images of any kind? if under 18 that likely flagged something honestly
No?
I'm going to make a new account
Am I going to be banned on the new account?
I Also had to Make A New Account Cause I Deleted Something In Sora 2 that Broke Everything SO Yeah
Hey guys i created a Securelocker (Bitlocker) with Codex but i have a bug and Codex cant fix that when i want to type password (example: i want to clear the logs i need to type my account password in the textfield and when i type the background is turning to complete blue
I try 10 prompts by Chatgpt but codex cant fix that
example prompt from the 10: ||`Fix critical input bug in SecureLocker:
ISSUE:
Whenever I type into password fields (e.g. Clear Logs confirmation, Delete Account, or any secure input),
the entire app instantly turns blue/blank and becomes unusable.
The app must be restarted after this.
TASK:
- Identify what triggers the full-screen blue overlay on key input
- Fix the root cause (NOT a workaround)
- Check:
- global input event handlers (keydown, input, focus)
- overlay/loader/progress systems
- modal/backdrop rendering
- any state change triggered on typing
- Ensure typing in any input NEVER triggers a full-screen overlay or UI reset
REQUIREMENTS:
- No screen flicker
- No overlay triggered on typing
- No app freeze or blank screen
- Must work in ALL modals and ALL input fields
GOAL:
User can type normally in password fields without the app turning blue or breaking.`||
I mainly use GPT-5.4. Now that GPT-5.5 is out, I’ve tried it as well, but it still can’t handle it. I usually use the reasoning level on high, and sometimes medium just for UI changes.
How to set GPT 5.5 to pass Context limit? its like defaultting to 256k
before with 5.4 i set to 750k in config.toml
isnt 5.5 400k in codex max?
Ask gpt (:
Damn i hope they bring back the 1m in codex 256k sucks for big codebases
haha
anyone have any alternative openai support channels? stupidly on my part, I didnt create a non SSO admin for my Team, and of course the SSO config expired. the email support was unhelpful to say the least and for several days now has stopped responding entirely. I have no way to disable SSO or login for my entire org.
Why my image generation look blurry, how to get the best resolution on image generateion?
Hi, is it no longer possible to use DallE in ChatGPT?
Older models through custom GPTs would retain their DallE generation
I tried generating DallE with a custom GPT and it didn't work
Is there any way I can request the data for the banned account be deleted?
My appeal was denied, so it isn't as easy as it would be normally
Anyone found a tool that alerts you when you start 'drifting' into AI rabbit holes? I keep getting sidetracked and need a way to stay locked in.
how do i fix chatgpt getting stuck at deep researching everytime
@tame cliff nice job getting hacked
I need help to export all my sora content.. I followed the instructions on export data, I received an email that says it all is available, but all i get is this , no files
Again, Sora 1 stuff is not included. You have to use one of the tools I recommended yesterday.
And yeah, I know it sucks that it has to be this way and not just through the official way.
hold on, i might fall asleep before I read those
thanks, hope this works 🙌
Sweet!
but this takes only images and not videos ?
Hm, I'm not sure actually, I assume both.
hmm was pretty sure I had more than 418 things there, but will see
Alright let me know once it's done 😄
i can see it downloading videos atleast
That's good! I do hope nothing misses from the end result.
it got everything from 2026 it seems, but i remember there was issue earlier where i scrolled far back it said something went wrong and could not go past that point, so there might be stuf missing from that point on,, not sure how tht can be saved, this is a problem on theire server end
maybe this mrror mode will work
with htis I atleast got twice as many
17 failed, these must probably be moderated content that triggered some failsafes i guess
Hm, might be possible. Did you at least get most of the content you wanted, or are a lot that you wanted still unable to be downloaded?
yeah, i'm sweeping the trash as well. just because why not 😄
it seems to maximum que 500 items, not sure how that affects it tho
how do I get gpt to create 10 images individually with 1 prompt? cause everytime I ask it do that, it just bunches of 10 different images into 1 image generation
anyone having trouble with signing up for a paid ChatGPT plan?
I'm trying to reactivate my Plus plan by redeeming a free 1 month offer, but it's just stuck on this screen:
^ yeah same issue without redeeming anything
anyone else noticing after this update that their pinned chats are gone?
ChatGPT and Codex are broken for me and everything under chatgpt.com/cdn is 404
yo @gilded meadow if you ctrl+shift+r it's fixed now in case you haven't yet
oh yeah just checked, thanks for the heads up!
slow for 5.4 today
Why its happening ?
Not sure how to fix it (other than using a different method), but it seems to be an issue mostly happening to people from India (or with an Indian card rather, not necessarily people from India)
From what I've noticed
I tried with another card too. Not working
Which is bound to what country?
India
Yeah, might just be something with India in general.
How to verify it now ?
What's the solution
^
Well, there are employees here, but for support, yeah, you have to go to website.
Apologies for my ignorance, but what exactly does this mean "Update: GPT-5.5 and GPT-5.5 Pro are now available in the API."? I am using GPT 5.5 Pro on my desktop using the app...
api is a separate paid service which allows you to use openai models in your own apps, games, etc.
Ah ... thank you, understood.
is tool calling broken? I'm on the plus plan and as of today I cannot get chatgpt to look up anything even when I expressly tell it I want it to search for the results.
under heavy load, #codex-discussions
Could be heavy load, but from the screenshot alone I wouldn’t assume that. It could also be a local app/runtime, network, VPN/proxy, or auth/session issue. Logs would tell more.
Very often it’ll say network disconnect though my net work is working fine and then I have to restart the app like 3 times any fixes?
How to upgrade the plan! any one facing same issue?
My admin just updated my Codex plan to Pay-as-you-Go. When I sign in to my account in my browser, I see: Your plan does not impose Codex rate limits, but no mater how many times I /logout and then codex /login, my terminal still shows: You've hit your usage limit. To get more access now, send a request to your admin or try again at May 21st, 2026 9:50 AM.. Any assistance on how to get my terminal to recognize the new plan, would be appreciate. Have opened new terminal sessions.
I am still having trouble downloading my export. I tried again today (for the third time), and then received an email a few minutes later saying the export is ready, but when I click on the link, it says no export is ready.
Same error for me right now.
I managed to use soravault from github. Effortless. Now I have peace of mind.
Clips: when you go to sora.chatgpt.com/exports/me (make sure it says "exports" instead of "export") > click "export" and wait for a little bit > then after you get a mail go back to the Sora.chatgpt.com/exports/me [manually. The link in the mail doesn't always work] and you'll get a list with zip files that you can download. This will contain both posted clips and drafts. (Posted clips will be in the spot where it'd be if it were still in drafts, it'll be all in order)
This is different from the "Export Data" button in your settings. Which will only give you GPT related stuff.
Oh nvm.
🙏
Is there a way to enable timestamps for the AI? It's annoying that I have to keep telling him everyday what the date is.
Hi everyone,
I’m trying to understand if this is affecting other people too.
For me:
- normal chats upload fine
- GPT chats on web upload fine
- GPT chats in the macOS and iPadOS apps fail with a 500 error
This happens across multiple GPTs and file types (JPG, TXT, PDF, DOC).
Is anyone else seeing this?
and it has been happening for about 4 days.
It should know the time amd date. Mine does.
How do I generate images in 2k? Just prompt "2k" in the chat? Or do I need api?
Hello, is there a way to move a subscription off of gplay?
I don't want to manage it with google play.
I'll let it lapse and then resubscribe on the web
Hi everyone, I could really use some guidance.
I recently lost access to a workspace called “CloudSoft Workspace” that I was a member of. It was suddenly deactivated, and with it I lost access to all my chat history and important personal projects.
I’ve already contacted OpenAI Support and my case was escalated, but I wanted to ask here in case anyone has experienced something similar.
Has anyone been able to recover data or regain temporary access to a deactivated workspace? Or is there any recommended step I should take while waiting for support?
This is really important to me since I had a lot of work stored there.
Any help or advice would mean a lot. Thank you 🙏
My image style suddenly changed drastically I mean I've been with chat GPT for a long time now. Pretty good at making the art work but no matter what I can do I'm getting artwork that looks wet and plasticky.
The output right now has that:
- thick edge definition
- chunky brush-like micro-contrast
- over-baked texture pass
- pseudo-oil-paint sharpening
It’s reading like “digital painting pretending to be realism.”
Pronpt Remove artifacts, make the picture smoother and less chaotic
i was use another but this one is more shorther let try then ^^
Indeed is most better !
i had this same exact issue i conntacted OpenAI they said to contact bank, bank said everything is ok at the end i ended up using my amex with apple pay and it worked then i went in to billing after creating it and changed the billing to my real account and it worked. I think this is a issue with stripe
I keep getting this error anything anyone tries to sign in
"Authentication Error
An error occurred during authentication (get_chatgpt_account_error). Please try again.
You can contact us through our help center at help.openai.com if you keep seeing this error. (Please include the request ID 9a692869-afe7-42cc-8a25-f8fe05f2937e in your email.)"
For browser:
Open browser > settings > find "site settings" > then "all sites" > search for "OpenAi" and "GPT" > delete data > then reopen the browser and try again.
For App:
Go to the homepage on your phone > hold your finger on the GPT app icon > select "App info" > find storage > "clear data" > reopen the app and try signing in.
(Don't worry, none of this has an effect on your account data)
Follow those exact steps and it'll be fixed. (You're not the first person I've helped with this)
can anyone suggest me an AI model (can be paid) that can generate me realistic model like images for my clothing brand, I wanna use some images for my website and some for model photoshoots etc...
GPT-Image 2.0?
is it better than nano banana?
I think it is, your taste may differ though, try it out is my best advice
I did try it and it's honestly pretty good, there's an app called "higgsfield ai", It has both the api's - the nanobanana and the gpt, I'll test both out and check which one's better. Thanks for letting me know!
can't sign into codex.....
it just sits on this bloody screen
and I am assuming there's no other way to sign into the app on mac
use your email and password if you set one, otherwise use sign-in with Apple
joy
doesn't work
how did you sign up originally?
through my gmail
this has to be a joke.....
I can sign into Google's Antigravity just fine....
what the hell are openai doing
FINALLY
it just worked, for some reason....
I was sat there for 40 minutes trying to sign in
Hi everyone,
I’m trying to understand if this is affecting other people too.
For me:
- normal chats upload fine
- GPT chats on web upload fine
- GPT chats in the macOS and iPadOS apps fail with a 500 error
This happens across multiple GPTs and file types (JPG, TXT, PDF, DOC).
Is anyone else seeing this?
and it has been happening for about 4 days.
the macOS ChatGPT Desktop App is working fine for me
Could you tell me what to do? This situation has been quite frustrating for days: both on Mac and iPad, the apps won’t let me upload files, but only in custom GPTs.
are you using this version of the macOS app?
yep
yep
can you walk me through what you do after opening the app, where it's failing?
Sure. I open the ChatGPT app, open any GPT, start a new GPT chat, click the + upload button, pick a small file, and then the upload fails with a 500 error.
It happens with JPG, TXT, PDF, and DOC files. Normal chats upload fine, and GPT chats on web upload fine. The issue only happens in GPT chats inside the macOS/iPadOS apps.
I just did that and it worked first time
I'll show you what happen to me
ok, I am waiting...
you see the gpt for swedish, the file, and the error
but if I do the same on line via browser it works
that's very odd, I think you might need to contact support at https://help.openai.com
via browser
Yes, I’m trying, but I’m only able to speak with a chatbot and it’s quite frustrating.
ask to be escalated to a human
ok I'll try and then if you could, I'm working as well so I can't stay longer here. But thank you by now
I'll come back to you
I am the sole owner of a new Business workspace (under 30 days old). I accidentally merged my personal account into it, and now I cannot leave due to the ownership lock. I have redeemed my student credits which require a Personal workspace. How do I create a new personal workspace in which i can use my codex student credits?
hi, can someone please help me?
i tried to log in with another pc, and now my codex looks like this
Hi! My account was suddenly downgraded from pro to free plan, what should I do ? It was working 2 weeks well and now just suddenly happen
how that can happen very crazy !
my account same downgrdaed from pro to free
anybody here having the issue on the api site for openai api, where everything is just frozen??
Is there anyway where you can select Google file docs instead of using the Google drive app and trying to find the right file?
I pay for plus, and was randomly hit with a "message limit" and I have never once seen this before since paying. Was I downgraded and don't realize it?
My accont downgraded from Pro to Free, what happened ? no email/notification received about that, it was just one week old
hey guys, im relatively new to the ai space and im looking to learn alot more about ai and its different uses. what would you guys recommend is the best way to learn the ins and outs of ai and learn how to incorporate it into daily tasks and jobs
Hello, anyone having this error:
openai.BadRequestError: Error code: 400 - {'error': {'message': "Invalid value: 'gpt-image-2'. Value must be 'dall-e-2'.", 'type': 'invalid_request_error', 'param': 'model', 'code': 'invalid_value'}}
When trying to edit with gpt-image-2 via python library?
are you sure you are running an up to date lib?
What?? I didn’t even use this device today…
What productive things can I use chat gpt for
My Google Drive folder is suddenly not syncing, Pro is supposed to have higher file limits in projects.
It’s treating me as a free user
Good evening, could you tell me where or which system is using GPT 5.5 to develop systems and even games? I saw a post on Instagram about incredible development with GPT 5.5 and I'd like to test it with my ideas. I already do development, but I think it would help me put some things I have in mind into practice. Since the space for women in programming is still small, I'm working on learning both manually and with AI, and the best tool is GPT. Do I do this using the Codex program?
Hacked
i have a problem with my phatgpt
because
its food tech was due after going to tietz crib
take melatonin tablets
consider changing your password and checking your devices for sketchy extensions then
Hi, I’ve been experiencing an issue where my ChatGPT app shows I’m using GPT-5.5, but when I check the model info on the web version, it shows the responses are actually coming from GPT-5.4. This has been going on for over 34 hours now.
I first noticed this because the response style and tone felt noticeably different from the previous days. I work as a translator, so I’m quite sensitive to shifts in language patterns — the change was obvious to me. I then checked the web version and confirmed that the responses were indeed coming from 5.4, so I’m confident this is not just a display error but an actual routing issue.
I’ve tried switching models and starting new conversations, but nothing works — it keeps routing to 5.4 despite showing 5.5 in the app.
I’ve seen a couple of other users report brief moments of being switched to 5.4, but they were able to switch back immediately. In my case, it’s been persistent and I can’t get back to 5.5 at all.
I’ve already submitted a support request via email and the help center, but haven’t received a response yet. Has anyone else experienced this, or does anyone know of a fix?
I think all OpenAI engineers are messed up. Today my friend tried to buy a subscription using PayPal, but the card got declined before even reaching the PayPal page. What kind of ridiculous payment gateway have they added?
openai never supported paypal and is using its competitor, stripe
Does the pdf viewer in the ChatGPT app for windows also not work for you guys? Also why do I see a blue circle with certain chats while no image or anything is generating?
@green summit any news on the bug?
Why? What Happened?
gemini forever
I subscribed to the gptplus, the payment is successful but it's not reflecting on my chatgpt account. Still in free version. Can anyone help me?
It probably already has been updated, but it may take a minute to refresh. Signing out/back in may also help. (For future reference)
Did someone stumble upon such error? If I make it from audio, the message is there, but it is not printing. Same with all chats in history.
Firefox Win11
Ok restarted and it started to work... But veeeeery weird. Thanks!
Does anybody have a usefull database / tutorial that are helpfull in creating better, more concise and concrete prompts? trying to work more efficiently
UPGRADE TO PRO 20x BLOCKED
Anybody else having problems upgrading from Pro 5x to Pro 20x? I've tried for two days and OpenAI Support hasn't been able to help yet.
I've tried from multiple different devices, desktop browsers, mobile, incognito, log out, log in, etc., etc., etc.
how do i fix this and why is it black now (it was just darker earlier)
i tried refreshing browser, website, reset cookies
troubleshoot mode doesnt fix it either
Yep, and my 20x subscription suddenly disappeared. 10 days early it vanished, it doesn't even show as being expired in my Google play account it's just gone completely. Basically scammed out of $66 dollars
Eh, what in the world lol? Why is conversation not found gimme my feedback bot back
Nevermind it’s back a few seconds later
not what you asked for but if you have a chatgpt audit your prompting style that you give it often, it will tell you what you are doing wrong using examples from your conversation history. just prompt it for that
i've added my card for plus subscription but i got mail that payement declined i tried to update the payement method but i'm not able to see the upi(india) method available earlier it was i'm only getting an option to add new card and even not able to use my existing card!
they are very strict with payments, and this comes as a problem in some countries, particularly in India
not much to do other than use a different card
you can also try to reach the support on help.openai.com
Sometimes payment through the mobile app may be accepted when the same payment method through a desktop wasn't. Otherwise, another method is the only option, and reaching out to support on help.openai.com is a good idea.
thanks idk y auto pay failed i'll try adding different card lets see!
For anybody else that is experiencing upgrade errors as shown above, you can work around the OpenAI Bugs by:
- Cancel your subscription
- Respond to their marketing email to Upgrade and you can buy 20x and unlock your account
OpenAI Support was non-existent after asking for a screen recording and later an HAR export (see screenshot).
Despite acknowledging urgency, they provided no help or guidance. It's been 3 days and still no response despite multiple outreaches.
For as much as I love OpenAI / Codex, Open AI Support Grade: D-
Does OpenAI accept PayPal for buyign Tokens ?
no
What payments do they accept
credit cards
you can also subscribe via the apps on the iPhone or Android
debit cards also work from some regions
I don’t wanna subscribe I wanna buy API tokens
credit card or debit card then
also depending of the model you need to verify yourself with a gov id
I have a MacBook Pro M4 with 24 GB of RAM and a 20-core GPU. Which model can I use for local management of tools, such as a computer?
ouch, that's not much memory
🙁
It's sad, I know 😄
I'm currently using gpt-5.4 fast and it's working fine, but I'm looking for a fallback
you could probably run gpt-oss-20b
why did chatgpt just randomly drop a hindi word
why is chatgpt high
how can i jailbreak chatgpt
just use a local heretic model
jailbreaks don't really work that well anymore
My new agent.md uses so much context it burns through the five hour window in like 30 minutes.
How can I reactivate my chatgpt business? even when I've already paid for all of my bill, the workspace is still being deactivated
you're not supposed to use a bloated agent.md
It isn't bloated. The LLM just has a lot of forsight into what it is doing. alot of feedback loops, etc
but it's been tested, a very long agents.md makes a model perform worse
how many lines/rules are u giving it
About 35
I compressed them, they were 610 words, but I had them remove duplciates, etc and it is down to 360
that should be good
My rules are very effective in that they work and make better code, but they aren't always effeicient.
currently forking a mature project and trying to re-factor it based on my rule system. Solarflare was able to do it a bunch, why not.
This is just a hobby , so if it fails, no worries.
Did anybody else get a new ChatGPT AI Agent today? I've been using ChatGPT for 6 months and the AI really understood my project becuase it was there from the beginning. But today, a totally different AI personality showed up. It felt like a C student 10th grader. (I know becuase I was one back in the day). It appologized over 40 times and told me it would get better but it never did. It was just weird. I feel like the team I was working with all got let go and new a apprentice AI showed up. It knew nothing about my project and told me things is a way that I knew it was broken. Did this happen to anyone else today?
I noticed this today, see my post below and tell me if i explain the same thing you experienced. It's like my AI got laid off and replaced by an apprentice. Totally, different personality and approach, and lacked a lot of knowledge of my project.
you mean it starts talking a little differently?
I would make 5 new instances asking for just image generations, or so, and then see if the personality is still here. sometimes, at least they say, they're testing new personalities, but sometimes it gets bent on your existing cross-thread context and acts dumber than usual
whats that can u send link
a type of uncensored model
go find one on hugging face or ollama
Unable to generate images - anyone?
Yes, but not a little differently, a lot different. Its whole process of explaining things reverted back to an elementary, super detailed response. And the details are things that it already knows from the project and explains very elementary things. We were on a senior engineering level and it reverted to entry-level responses.
Ok...urrm... thanks for nothing?
Ponders for 100 minutes and comes back with "Stopped Thinking" every time (5/5)
How or where can I set up a rp with chat?
Hi, My chatgpt account was recently deactivated for Cyber abuse (I am not sure why, i had tasked codex to work on my codebase to look for vulns and fix them as well as work on a dynamic application security testing agent (DAST agent) and the very next day my account was deactivated. On appeal i was denied. anything i can do from here?
Do DM me or ping me please
I'm not sure our issues are the same though. In my case, GPT-5.5 thinking is consistently being routed to 5.4 with no way to switch back. The model itself isn't broken or lacking knowledge, it's just the wrong version being served. Your situation sounds more like a auto- routing/downgrade to a lower-tier model, or maybe a context loading issue. You could try asking it which model it is, or reach out to support directly.
Thanks for replying Clingstar.
Good suggestions. I did ask what model it is and it didn't change. I tried to contact support directly and it pointed me here. I'll see how my AI responds today and dive deeper into contacting OpenAI support if needed.
I just found a way to check which model is actually being used. Request a data export from Settings > Data Controls. In the exported JSON, look for default_model_slug and resolved_model_slug, if your default says 5.5 but resolved says something else, that should be a routing mismatch.
I'm going to have to give that a try. Conversation I just had with chatgpt told me I was on gpt-4, but there's nowhere to choose that. I've been using chatgpt for a project since last November. I'm guessing the change I see was a forced change to use version gpt5 -x.
It was just a very drastic and obvious change.
I'm having to retrain and fully second guess code that has been provided by chatGPT because I have found the new version making multiple oversight and leaving important code statements out of a revisioned file. I'm being forced to use the compare feature more often. 😆
I’ve been a movie critic for 3 years and over the years I have occasionally had people say my reviews feel AI generated even though I write them 100 percent myself so I check some of them and most of them come out human but a chunk of the AI checkers I tried all of them came out saying they were AI generated is there such thing as false positives because I write them all by hand here is my page on Letterboxd if you want to read them for yourself or Instagram my name is Two Thumbs Up Movie Reviews and More
hey i’m doing a side project with the projects option
it worked and it could read all my pdfs that i added to the data base but for like a day it is not reading nor utilizing the data within
it says that is a index error because the document was uploaded but the messages couldn’t be finded inside the pdf
is there any solution to this problem?
anyone else having issues with the ChatGPT desktop app (macOS) ? after each response the text input field becomes disabled , i can fix it by switching to another chat and then going back to the original, which seems to 'refresh' or something and then i can enter a prompt again. Not the end of the world but its pretty annoying
i've tried reinstall no change
the macOS desktop app is working fine for me
hm ok thanks for checking
No fix yet. Been happening since a few hours before they added GPT Image 2.0, and hasn't been fixed yet.
As per usual
so GPT-5.5 instant on API .. is it just same pricing as regular 5.5? and currently only way to call it is via chat-latest?
Is anyone else running into Codex workflow issues like these?
I’m mainly seeing:
- auto-compaction disrupting context/workflow continuity;
- queued messages overriding or interfering with pending plan-mode approvals;
- plan mode becoming hard to trust when messages are already queued;
- uncertainty about whether there is a safe sequencing habit or workaround.
I’m not just trying to complain — I’m looking for practical workarounds. Should I avoid queueing messages during plan mode? Start a fresh session before approvals? Keep plans shorter? Or is this expected behavior right now?
yes.
any workaround or feedback?
honestly I think it's the Codex program's fault, the model is great. Maybe try 5.5 as a model in VScode or another wrapper
you can use gpt 5.5 in opencode
I was wondering if the username from my old account will be available? If not how long should I wait for it to open up?
How do you deal with artifacts in Images 2.0?
idk use nano banana 2 gpt image 2 kind of sucks rn
they nerfed it a lot since launch
this is bug or feature? Why
Why,the “show in text field” of button has disappeared
Now uploading pasted content as a file loses a lot of information
And this feature has been removed; it’s obvious that the intention is to reduce the length of context information.
Can someone tell me if this is a bug or a feature?
This determines whether I will consider paying next
Double your normal Codex usage on the $100 / month tier until May 31, 2026. is the deal, does that mean it is equal to a 200month account until may 31st?
feature ig.
ChatGPT, as a text, has a word limit.
At some level, it does accept text in chat feature too. But when it goes beyond maximum threshold, it would process the file only.
This is definitely a bug because the original word count limit is inconsistent with the current limit on the length of the automatically converted text
"If you paste more than 5k characters into the composer, ChatGPT will automatically convert the content into an attachment instead of inserting it directly into the text field. This keeps the composer cleaner and helps prevent large pastes from consuming your entire context window.
You can still move the content back into the message at any time by clicking on Show in text field to convert the attachment back to a direct paste."
So, how many characters is the context limit for a single dialogue input box
Currently, I'm trying to automatically convert it to a file when it reaches about 25,000 characters
But I can paste 20,000 characters twice(about 40,000), so the context length of a conversation is definitely not 25,000
This leads to a problem: I originally intended to directly paste 25,000 characters, but now I have to paste it twice; otherwise, it will automatically convert it to a file for me
The reason why I don't want to use file upload is that the current mcp tool for file upload parsing definitely has poor performance and will lose my text content
So is this restriction really a feature? The back-end service can accept a maximum of 25,000 characters in a single conversation, but the front-end restricts pasting to two separate times; otherwise, it will force my text content to be converted into a more error-prone file for upload
What do I do to fix this error?!
It's in old and new
When I use thinking the first image is fine subsequent regens show this
its a new feature and was listed in changelog 1-2 weeks ago
max context window is plan based
So in the future, I won't be able to get 25,000 context Windows in a single conversation?It is still available now because the backend service is still in gray-scale testing and has not been fully released yet
I honestly think it’s a good change tbh. Huge walls of pasted text can clutter the conversation and eat context really quickly. Converting it into an attachment keeps the chat cleaner and probably helps with context management on the backend too.
I get your point about file parsing being imperfect, but from a usability standpoint, huge pasted blocks directly in the chat can become messy really fast.
It's not good. Using the mcp tool for file upload will cause a significant decline in the model's performance and the problem of reading fragments
And this is unfair to users
One can no longer choose what they want independently
An accurate chat should solve the problem directly and not produce too long contextual text, but due to the low quality performance of the MCP tool for file uploads, people will need to have more conversation to correct the model reading only part of the uploaded file
If you want to use better context management, you should use codex
Hey everyone 😅
I accepted a ChatGPT free trial a month ago and today I got charged for the subscription renewal without realizing auto-renew or something was enabled
I’m trying to get a refund ASAP but support has been impossible to reach
Does anyone know the best or faster way to contact OpenAI or get a refund?
you signed up for a subscription, if you didn't want it you should have unsubscribed before the month rolled over
Yeah, I understand subscriptions usually auto-renew but I just didn’t realize it was still active, and right now I’m mostly trying to figure out the fastest way to contact support and request a refund
Also, the billing page has been quite confusing because it won’t even let me remove my payment method then and now
you can get support at https://help.openai.com
The issue is that the official billing articles specifically instruct users to contact support through a chat widget that currently does not appear for me at all
That’s why I came here in the first place 🥶
the chat bubble on the help page is in the bottom right hand corner, you need to click on it to make it bigger and type your question/request
i dont know. now that i checked its very old change by now https://help.openai.com/en/articles/6825453-chatgpt-release-notes#large-pastes-are-now-handled-as-attachments
That’s the problem
The bubble literally does not appear for me at all, which is why I got confused and came here looking for help....
have you got another browser to try?
You're right, I'll try it
I tried again from my phone using Chrome and the widget actually does appear there
But as soon as the conversation continued, it asked me to log in, so i logged in and the widget completely disappeared
Then I refreshed the page and the bubble appeared for like half a second, i tapped it, it opened, and immediately closed by itself
Now it just keeps repeating that same loop every single time
refresh, bubble appears a sec, i tap the bubble and instantly closes
use another device
use your computer and try Firefox browser
Same
ok, well as a last resort try emailing support@openai.com
expect to wait days for a reply though
You could’ve started with that email from the beginning lmao
But thnx anyways
it's not meant to be used as first point of contact, the help website is the official channel
chatgpt still so buggy for me... still creating pics for me when i just wants texts and no pics 🤦♂️
Hi, can someone from the openai support help me with a problem? or someone here can send me a contact to talk to?
whats this?
Tick?
yoo guys, i was tryna verify on https://chatgpt.com/cyber but when i scan my id its always denied, does anyone have the same issue?
try another ID. What country are you in and what ID have you tried?
indonesia, mann i've tried my id, my dad's and mom's id
and still error :'>
use your drivers license or passport
you can't use your parents ID, the ID must belong to you
im not trying it yet, i just have no idea how this issue could happend to this such a big tech thoo
it's a simple process to verify you for cyber activity, I used my drivers license and it was easy
imma try it out, btw u do cyber security too? what do u do bro?
I do a lot of stuff, see my bio for details
Ur Cool bro
why Venezuela is still blocked—we're already the 51st state 🙁
The image tool is annoyingly aggressive
this is a lone star tick, I do hope you were not bitten
Yeah.. be careful with those.. they can carry many diseases.. and that particular one one that can make it so you can basically not eat any red meat anymore and possibly dairy and some medicine aswell. It makes you allergic.
Tell it that.
The tool it uses to get images from the net is images_group.
You can tell it you just want text no pictures, and tell it not to use images_group tool can help if there's confusion.
You can tell it what you'd say if you did want it to create a picture or wanted it to show you pictures from the net, and that can help even more, because then it has reason to believe you will ask if you want it (supports that you don't want it if you don't ask).
It doesn't listen
Huh. Would you make an example chat showing the problem and share it?
If you're interested, here's a test chat where I explore it responding with a semi-casual flexible ruleset about images, mark where and how it may get confused (one example), then discuss with it a firm ruleset a person could use to be very clear that images are not welcome, unless they are that one input you're asking.
https://chatgpt.com/share/69fd83d6-dbfc-83e8-bfb8-260f62bb2611
Lemme take a look
yeahhh ikr😑😑
how i tell then?
@vocal vector dm🙏
Like talking to a person. Just input it, same as you say what else to do.
i have many times… still not working
oihyopyh
Want to share a short chat that shows the problem?
Here is a short chat where I explore how to make it work.
https://chatgpt.com/share/69fd83d6-dbfc-83e8-bfb8-260f62bb2611
still not working for me :D
Would you like to share a chat that shows the problem? A short one, so people interested can look and show a fix or help report a problem?
im literally saying no pic :p but its creating a pic for me
If you want help and fixes explored, share the chat?
bruh
Or make a new chat and show the problem. Good if your chat is long or personal
😭 wtv
im getting this error when im trying to upgrade my plan to pro?
(my bad if this is the wrong channel to ask for this)
<@&1150054613710557215>
Has anyone else had GPT-5.5 Thinking consistently routed to GPT-5.4 Thinking the entire time? Mine has been stuck like this since April 29, and I haven’t been able to use 5.5 Thinking at all. It’s super annoying, can’t believe they haven’t fixed this yet 😑
Reach out to OpenAI tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support?
honestly, this is likely their new prciing tiers
idc abt the pricing i just can't upgrade my plan
try upgrading from the iOS or Android ChatGPT app
Or into chrome browser then back to gpt
Assuming u habe andorid
Id you habe pasw
That do was nst work update it
can't make on app because "i made the purchase on another app" let me give you money openaiiii
What u building
reach out to support at https://help.openai.com
Or go to cursor meantime
yeah i did will recieve a mail in the following days
on claude actually
ok, hopefully they respond soon
its fine but opus get's lobotomized randomly
On claudes?
why do they not have support on discord directly lol
Claude code
yeah
but the rate limits i needed to upgrade to pro today but the site won't budge
It is a good idea to not rely on a third party for a matter like that
also... discord is not a good platform for that
trust me.. I made @burnt galleon
lmao
yeah ig but railway has really good discord support and it's quite fast on their server
Guys, i want to ask... can the output of GPT image (GPT image 2 or previous versions) be used for commercial use (print on demand, design for clients, logo for clients, stickers, etc.) even for free tier users?
Tried asked chatgpt and the answer is yes, it can be used for commercial use.
Typically yes. Check ToS, there's a link to Terms of Service in #server-rules for any specifics that apply to your intentions.
And maybe note that the rest of any commercial use chain may have their own requirements, as well as OpenAI's ToS.
I've read it. I'm just worried i might misunderstood about it, so I'm asking here to get the right answer.
You might want to consult a lawyer to ensure that for your situation everything's exactly right. We're community members here, even the mods are volunteers. We help each other as we can, but even as a Guide, I speak only for myself.
If you're concerned, I highly recommend you consult with a professional about the specifics of your situation and anyone else involved and their requirements.
Surely the mods have closer connections with the higher-ups than community members, unless they aren't that involved besides on Twitter?
I want to volunteer to be closer with them too 😂. Maybe on a different account sometime
Oh, i think no need to do it, i'm here just curious and wanna know about it.
And sure, thanks for the answer 🤝
just checking if there is any Business or Enterprise user manage to create new skill with the built in skill-creator or upload a new skill successfully? thanks for responding.
hey guys i bought business plan with a promo code and it says that 1 seat is free and i only use one seat and in the billing page it says that i will be billed only for the seats that are assigned which is 1 and so am i going to get it for free?
who are interested in bittensor?
are you going to spam every channel?
no, i wanna ask to everybody as soon as possible.
only post once in one channel, what you just did is considered spamming
if you are not interested in my question, pls ignore that.
the correct channel for your question is #ai-discussions
thank you
hello guys, I want to subscribe to chatgpt go. But only if I get more usage for codex and the codex api would be great
Idk why but gpt chat keeps toggling the image generation thing and I don't want that
or chatgpt plus would be an option if i get an API
but I don't know it is a seperate product to codex
Edit: Partially solved, after over 2 hours the chat history in the app has started to repopulate itself, not fully yet though. Original issue: A lot of recent chats in my chat history just disappeared (I'm using the Mac App), anyone got a solution to get the chats back? All chats are visible on the website. After logging out and logging into the app again now the entire recent chat history is empty. Would greatly appreciate som helpful ideas.
Anyone else having the issue that OpenClaw is breaking all the time. I tried to add more models to the list. It's really annoying and I'm not productive that way
The help files don't clarify what happens for Business users who fail to maintain minimum seat requirements, but the minimum allowed is 2 unless you've signed up for Codex only; Codex only has no minimum seat requirement, however Codex is 'pay as you go', and I suspect does not allow free access at all.
https://help.openai.com/en/articles/8792828-what-is-chatgpt-business
I expect that something may happen to your account if you have fewer than the required seats, so you might want to discuss specifics with OpenAI tech support,
https://help.openai.com/en/articles/6614161-how-can-i-contact-support?
We're community members here, similar to yourself, we can discuss what we're aware of to help each other, but we can't make stuff happen to solve problems or clarify like an OpenAI tech support person could discuss or help with.
Hi chatgpt 5.5 is reliabel for learning ?
Something is wrong with the Google connector again
not if it's your only source, but it's fairly useful for breaking down complex topics so they're easier to understand
chat teaches me better than my teachers or even youtubers XDDD
today i got unknown error when attach .md files. i usually do those just fine. attach image still work.
How do i download my sora videos cause it is banned now the app isn't working anymore
The only way to get your stuff is to go to sora.chatgpt.com/exports/me > click the button > wait for mail and manually go back to sora.chatgpt.com/exports/me > list of zip files with all most of your clip. [Posted clips will be at the spot it would've been if they were still in drafts, it's all in order]
[some copyrighted stuff might not be included].
Also:
-
only for Sora 2 (Sora 1 stuff is not included).
-
this is different from the "export" button in GPT or (when it was still available on the web) Sora.
-
May take some time depending on how many clips you have, but it usually doesn't take that long.
And one more thing, it'll also put your prompts in* txt files aswell.
Still works
Hi everyone I got a question:
I’m subscribed to the $100 ChatGPT Pro plan, which gives access to GPT-5.5 Pro / Frontier Pro in ChatGPT web. In Codex, my model picker only shows GPT-5.5 with intelligence levels like Low, Medium, High, and Extra High. Does Codex actually use GPT-5.5 Pro / Frontier Pro for Pro subscribers, or does Pro only increase Codex usage limits while the available Codex model remains regular GPT-5.5?
GPT 5.5 Pro is not available in Codex (yet).
The $100 subscription is basically just a limited version of the $200 plan, so less usage, but that's it. Everything else should be the same as the $200 sub.
So Im not able to get the pro reasoning in codex right?
Hello, I'm having trouble creating my companion in Codex. When I try to create it, my chat log disconnects every time. Does anyone know why?
Hi, I am facing this problem of "Suspicious activity detected". It's been 2 days so far.
Yesterday, I logged out of all devices and changed the password.
have you got 2 factor authentication set up?
Yes, but turned it off after a while
turn it back on
done, now back to waiting randomly?
if that doesn't solve the problem then you need to reach out to OpenAI Support at https://help.openai.com
It's either AI chat or Q/A. No real humans to contact there.
ask to be escalated to a human
It is a bit inconvenient, but it worked, and I got to talk to a human being via email now 🤡 thanks!
Does it must be on github or hugginface is also possible for codex oos?
it is probably a good idea to changle your password and also enable 2FA with an app like google authenticator, not just via email
also.. might be a good idea to review your emails regarding logins and do a check on your accounts, just in case.. it will not hurt xD
Actually I am trying multiple harnesses flavors; codex, githubcopilot, claude code, antigravity and others in the same project. I wonder if it is necessary to be redundant by creating .codex .claude .github....folders and files or is there an optimal universal way? heard of .ai + "symlinks", FYI it is for project scope. TY (reposted to the correct channel)
general instruction for an AI is important to mak better answer ?
How to generate multiple photos at once?
open multiple tabs
Ask it. But enable thinking mode.
it doesn't work
Free or paid user?
does anyone know in how long does DALL-E-3 deprecate?
Some time today.
No exact time
aw i was expecting it to be 2h ago since that's 12mn 12 May in San Fran
nothing in the announcement either right..
how do i get acces to sora 2
yeah, actually, how do I also get access to it
I tried to access from the website and got a ton of errors, prohibiting me from actually accessing the site
Sora 2 has been discontinued
wow ok
is there any other way of creating videos for free which are over 60 seconds long
for free
yes please, this
Thinking mode is only available on Plus and above
how much deep research usage with gpt plus?
I got a pro subscription recently and I just saw this student promotion for $100 of free codex credits. I'm not entirely sure what $100 dollars of credits approximates to in terms of how long it lasts versus the monthly plans. I had a question about if this would work for a free plan. I personally never use gpt enough for it to exceed the daily or weekly limits, so would this $100 dollars be more effective stacked on top of my monthly plan, or would it be better for me to cancel my monthly and use the credits on a free account till they run out?
Nvm. I thought this channel was for direct help.
for people using the personal ChatGPT Pro ($200) plan heavily with GPT-5.5 Pro:
Have you ever hit a hard limit where access was blocked until the next billing cycle/week, or are the limits mostly rolling cooldowns (e.g. wait a few hours and continue)?
because i have been on the business plan before (had a seat) and once the limits were reached you had to wait another month. i was wondering if it was the same for the pro plan as im thinking of upgrading
I use it a lot and I never reach any limits
I used to be on the $100 plan, and some times I would reach image generation limits.. now I don't reach it even if I try...
one really gotta be purposefully wasteful to reach the limits on the $200 plan
like.. give codex enormous and impossible tasks, bloat it with as many skills as you can.. only use the most expensive model in max reasoning length...
doing that is jsut dumb.. if you simply use good settings to a task and use it efficiently, you will never reach the limit
and "use it efficiently" im not meaning "economically".. I really mean.. use it in a way that will make you seve the most amount of time and effort, as in using it effectively..
because using it wastefully will not give you better reuslts.. it will be probabily worse
5.5 Pro right?
Im using mostly 5.5, but I usually don't use Pro because it would not be an efficiant usage
I don't want it to write a paper for every dumb question xD
on chatgpt, I usually go with thinking in medium effort, it is enough in most cases
otherwise, it will jsut keep thinking for a long while for something that don't really that much thinking.. which jsut makes it take more time to get the answer
in Codex, I usually have it on medium effort at 1.5x speed (which uses 2x tokens)
or, in more complex tasks, 1x speed in xhigh reason effort
I also have made a preset agent with Codex 5.3-spark for "explorer" tasks, that involve finding stuff or reading lots of content, because 5.3-spark is absurdly fast
Hey i wanted to ask how to use sora 2
Why can't I sign up for the ChatGPT Plus trial even though my Visa card is active and has sufficient funds, and my account has never used any services before? I registered in my country and didn't use a VPN, my Visa card is from a local bank, and my personal information is correct, yet I'm still being rejected.
Hi, which plan would be best for a team or 3? Business or plus? Per month, do business and plus provide the same codex usage? What exactly is different from plus/business? If you add people to your workspace in business, would they consume the same codex quota?
business is a business plan, terms and features. codex quota is not shared, same limits as with plus
It doesn't exist anymore since April 26
So each member in workspace gets their own limit correct? How is that limit compared to plus?
Does anyone know why Codex is so laggy UI wise, I am constantly getting wait to respond. Meanwhile claude is perfect
yes and same limits on both with codex
Does anyone know if 4o ImageGen is completely discontinued? I was able to use it for a while after they stopped making it the default by having it pinned to my sidebar, but today I went to make some pictures with it and it just used the default image engine
Hi everyone, I've been using a chat I created with the Study & Learn feature all day for a math test I have tomorrow. I accidentally closed the chat and it disappeared from my chat list. How can I find it again?
You may want to check the API/playground. https://openai.com/api/pricing/ shows a price for images with 4o at the very bottom, the calculator.
I don't use the API, so I'm not sure if that's still offered or just a relic on the calculator.
I am unsure, but this page may link to that part of the API at the top (I haven't made an account so I can't easily check). https://openai.com/index/image-generation-api/
Explore OpenAI API pricing for GPT-5.5, multimodal models, and tools. Compare token costs, realtime, image, and video pricing, plus service tiers.
Close the chat? Do you mean you deleted it?
And unless you named the chat, and if you do remember it, you can search it up in the search bar and it should pop back up. But if not, you’re gonna have to start a whole new chat. Unless you archived it.
hey everyone I'm runniing Codex on windows it's constantly crashing even when first opening it say's : Not Responding, how can I fix this or update it?
I keep getting this message while using chatgpt for chats or images: Streaming interrupted. Waiting for the complete message... I tried clearing cookies, refresh only works temporary.. What can I do?
When I told chatgpt about it it stopped doing it almost as if it decided to annoy me 🤨
I’m a little confused as to why my account has been suspended
did you get an email?
Yeah, I checked the policies and am still trying to figure out where any violation is..
the email should say
But I don’t understand why
does it give you instructions to appeal?
Yeah, but how can I appeal if I don’t know what exactly the reason was?
I don't know
I’ve been using it in correlation with my livestreams to showcase what is happening and what was done on stream, create images based around that.
I didn’t see anything prohibiting me from doing such activities
I don't know, something has been picked up though
That’s why I’m confused, as I also paid for more abilities
you could try contacting OpenAI support and ask what you did wrong
I tried with the bot in discord but it’s not working for contacting them drop down
no, you need to ask for support at https://help.openai.com
Thanks for that @charred parcel it does state this at the top of my screen so… yay! I just don’t understand, I got a onetime code when I was asleep which is strange because the email I use is a personal one not connected to my streaming for security reasons.
try emailing support@openai.com
i just looked at the timestamp of the email... (i was asleep when mt account was deactivated)
Anyone else having sudden trouble getting gpt image 2 to follow the prompts and espec ially, the reference images ? it used to be insanely precise and now it only very loosely follow them, creating something quite different and not consistent ?
Yes, I'm experiencing this right now
dang
it's quite clear since i've been running pretty much the exact same prompt for a while with different pics of cars and the difference is obvious
started happening yesterday or something
yep it's like doing an entirely different car now lmao
try turning memory off and/or getting out of the project if you're in one, seems to work better when i do this, it's probably overthinking/context pollution
why is my macOS desktop GPT app using 36GB memory 🤯
suspensions are delayed to prevent probing
Use "Thinking Mode" while asking it to generate an image. It will use reasoning and follow your instructions better for the image. [Which is a new thing they've added since image 2.0]
(Or should anyway)
Thinking mode (for paid users, not sure about free) will also allow you to ask for multiple (separate) images to be generated at once.
i niticed my email has a onetime code for my email at an earlier time that i was asleep as well, so im still confused lol
never used it outside thinking, even used extended but it just overthinks forever
I this a place to ask for some help woth openai and clawbot issues?
Here, #ai-discussions or #1070006915414900886 .
you are not the only one having this issue, seems to be a problem on OpenAi's side, try again later
Is Codex Prop part of GPT Pro?
Codex usage is part of ChatGPT's subscriptions and your usage limits on Codex depend on your subscription
For browser:
Open browser > settings > find "site settings" > then "all sites" > search for "OpenAi" and "GPT" > delete data > then reopen the browser > sign in and check again.
(Don't worry, this will not affect your account data)
Yes, this will 100% work. I've helped quite a lot of people with this.
Tried it multiple times.
Even other devices.
Did you delete both GPT and OpenAI in site (data) settings?
Which sign in option do you use?
It's odd. I've helped and done this method many times myself.
Specifically for authentication errors
Google.
My other google account seems to work.
And you are using the browser, right? Or is that somehow the app? (Possibly on a iPad or similar tablet like that)
Cuz if app:
For App:
Go to the homepage on your phone > hold your finger on the GPT icon > select "App info" > find storage > "clear data" > reopen the app and try again.
Hm.
I am using the browser and I tried 3 different devices on different networks.
This is not confusing AT ALL! wth!
It is something weird.
It started to show up since yesterday.
^ here's a message from me where I showed it getting fixed by following those steps in real time.
But you possibly don't have access to that chat due to it being in Archive, so here:
After clicking "Google" I stopped the recording for 2 sec so my emails weren't being leaked, but I clicked my mail and it immediately signed in (edit: as seen in the recording).
Anyway, following my ***exact*** steps will fix it.
And I attached these files:
This was recorded for Sora, but this is should be the same for (any) site.
Same steps.
Still the same error.
Hm, yeah that's not supposed to happen. I think your next best thing to do is to copy your ID shown in the error message and then go to the helpdesk and request some information on the matter, see if they can lift up the "block". (Ask to be forwarded to a human, might take a few tries)
Or wait a bit. But idk if that will fix it since it's been since yesterday already.
I can still not connect the codex app to my pc, why?
"Update the ChatGPT mobile app and the Codex app on macOS to try it out. Support for connecting your phone to the Codex app on Windows is coming soon."
Yo, can I report a member in this server for taking part in a raid on another server?
Or is it not possible?
Ok regens of images aren't showing up on my app only on the web. How do I fix it or how long does it last?
Hi for the new ChatGPT personal finance. Is the finances connector blocking other connectors from running in the same chat? I cannot seem to run finances + others in the same chat window. It loses connection to the other connectors
I've tried codex this week coming from CC and the permission model is simply messed up, if you try to work with MCP servers etc than EVERY single command requires an individual permission no matter how many times you approve it - what am I missing
it says this when i try to log out the hell is going on
Could possibly be caching something about your session, this looks like it's on android, is that correct?
Interesting, did it download anything when that happened?
i am on the pro plan of gpt and i can only see the extended option for the thinking time, isnt there supposed to be the 'heavy' option?
Hi so I was doing a huge pass on the online game I'm making for me and my friends and right at the end I reached the weekly limit for plus, but like.. it didnt finish 😭 i thought codex was supposed to finish the request before it cut you off
Did anyone get this bug? I can’t download my own AI image. It only shows a network error even though I have a stable Wi-Fi connection. What should I do?
anyone dealt with this before?
- 13 May: subscribed to ChatGPT Pro ($106 total)
- 16 May: accidentally upgraded to the higher tier, immediately asked the support chatbot for a refund
- got refunded $115.96 (the upgrade charge)
- but now my account is on the FREE plan, not back on the Pro I originally paid for on 13 May
so basically I lost ~$90 of unused Pro subscription. the refund only covered the upgrade, not the original plan that got wiped.
where can i get help for this?
i tried this keeps occuring
ok, send an email to support@openai.com
okay thanks
sending an email will be slower with response times than the help website usually is, but if the assistant bot isn't working then you aren't left with much choice but to email support
the refund i got for accidental purchase was through chatbot.
yup, thats the usual way to contact support
but it also revoked my pro plan subscription and after that it stopped wokring
i think i have got heard i got refund for 5x as well. thanks OpenAI ill subscribe back to 5x ❤️
Is there a way in codex to create an automation that will review something in an authenticated website? ie: a site that you need to login, or does the browser in codex that is used by automations always reset to a logged out state? I would like to create an automation that can check an internal website every day.
I'm sure it's been answered already, but my search didn't find it: Did something change in the last 48 hours that now prevents copying whole chats? For months I've been able to grab entire chat logs ussually in the 50k to 100k word range. As of last night I con only grab about 2000-2500 words, and I get a weird mix of the top few dozen lines from the chat and the last ~2000 words. Have tried in the pc app, Firefox, and Microsoft Edge with the same results
🤔
you're better off without Atlas, it is abandonware and probably insecure, use Safari or Chrome
Is there a way to delete a connection in the mobile ChatGPT app? I've removed my phone in the Mac Codex app but I can't remove the connection on my phone to reconnect. Now I can't establish a connection anymore :/
So my phone is gone from the Mac App but the connection still shows up on my phone. My phone says I can only disconnect but not remove a device. It sometimes connects but then all messages get a 403 Forbidden
(scanning the QR code again doesn't do anything)
I've tried codex this week coming from CC and the permission model is simply messed up, if you try to work with MCP servers etc than EVERY single command requires an individual permission no matter how many times you approve it - what am I missing
Hey everyone. I am in shock and need some help navigating this.
I received an email from OpenAI saying my account was permanently deactivated for alleged "Child Sexualization Activity". I have never produced, requested, or engaged with any content of that nature under any circumstances. The accusation is completely false.
I submitted a formal appeal clearly denying the violation and requesting a full review of my account logs. The appeal was denied with no detailed explanation.
I suspect my account may have been compromised, or that someone with access to my API key generated content without my knowledge.
I am now looking for advice on what to do next:
- Has anyone successfully overturned a denied appeal at OpenAI? What actually worked?
- Is there any way to reach a human reviewer beyond the automated appeal form?
- Would disputing charges with my card provider or filing a consumer protection complaint put any pressure on them to re-examine the case?
I am not willing to let this go quietly. A false accusation of this severity has real consequences. Any help or direction is hugely appreciated.
iirc and if you use the search for older post you can see that denied appeals have never been reversed. btw the support thing and appeal system has been an ai for a long time if it did not escalate to a human at any point
for account or api being compromised wont help as there are many safety options by them. if you are from eu i think you are still allowed to request full data export which would include chats
Anyone else encountering this issue with image generation in the app where the first image appears but regens don't appear at all only on web? How does that get fixed or is it a server bug?
After the reset in codex app, the remaining usage always shows 97% left instead of 100%. Anyone else has faced this issue?
have you got your pet running?