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.

No repository file, test, or entry point is identified. Start by comparing the behavior with issues #37518 and #38503, then determine whether normal conversation-history navigation can reproduce the restriction. Done should include a clear distinction between history access limits and model-usage limits, with sensible handling for a rate-limited history request.