Conversation history temporarily blocked by “Too many requests” during normal ChatGPT use
What issue are you seeing?
While using ChatGPT normally in Chrome, I received the following modal:
Too many requests You’re making requests too quickly. We’ve temporarily limited access to your conversations to protect your data. Please wait a few minutes before trying again.
The restriction blocked access to conversation history itself, rather than merely limiting model messages or usage. It interrupted an active workflow and prevented me from opening the relevant chat.
I was not deliberately sending automated, bulk, or high-frequency requests. The issue appeared during normal interactive use.
Why this is concerning
The message does not clarify:
- which type of request triggered the restriction;
- whether the limit applies to the current browser session, account, or conversation-history endpoint;
- how long the restriction will last; or
- whether ordinary navigation, sidebar refreshes, or opening conversations can contribute to it.
This looks similar to reports of repeated /conversations 429 responses and temporary conversation-access restrictions, including issues #37518 and #38503.
Expected behavior
Normal use of ChatGPT—including opening chats and navigating conversation history—should not trigger a temporary block on conversation access. If a history-related endpoint is rate-limited, the UI should distinguish it from a model-usage limit and apply sensible retry/backoff behavior.
Environment
- Surface: chatgpt.com in Google Chrome on macOS
- Subscription: ChatGPT Pro
- Screenshot: attached
- Date/time: 2026-08-15, Hong Kong time
Additional information
The problem resolved only after waiting, but it disrupts ongoing work and risks making conversation context temporarily unavailable.
<img width="1310" height="726" alt="Image" src="https://github.com/user-attachments/assets/0edb872e-6800-491e-9fc6-32f5af001540" />
4 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Also seeing this in a fresh browser session
<img width="932" height="342" alt="Image" src="https://github.com/user-attachments/assets/7c5484e3-0147-41f9-b0d3-ddb6f7fe7455" />
my list of chats isn't displaying so this could be chat enumeration hammering the server
Same symptom, but persistent rather than intermittent
Symptom
GET /backend-api/conversations?offset=0&limit=28&order=updated returns 429, so the conversation list never loads. The UI shows a modal reading: "Too many requests — You're making requests too quickly. We've temporarily limited access to your conversations to protect your data. Please wait a few minutes before trying again."Observations
Response time is ~800 ms, so this is an immediate rejection, not a timeout or a slow/failing query.
It is rejected at offset=0, i.e. the first page. This suggests the behaviour does not depend on the total size of the conversation history.
It also occurs immediately after opening a new chat. No automation, bulk operations, or rapid successive requests were involved.
Client asymmetry
Codex retrieves and displays the same conversation list for the same account without any problem.
Opening a conversation by direct URL works fine, so the underlying conversation data is intact.
Duration
Update — Aug 26, 2026
The issue is now resolved.
I disabled the GPT-related Chrome extension the previous evening. The issue
did not resolve immediately, but by the next morning the conversation
history was loading normally again.
One correction to my earlier observation: when I said no automation was
involved, that was accurate for my own actions. However, I cannot rule out
background traffic generated by a browser extension, which would not
necessarily appear in the page's own DevTools Network panel.
If anyone else encounters the same issue, disabling Chrome extensions may be
worth trying.