WhenResetHow to check your usage limits
When a model says you have hit a limit, the useful next step is almost never “search for the average reset time.” It is “find the meter that just tripped.” This checklist walks through the usual places to look and the questions that separate account windows from public announcements.
Sources reviewed:
Before you open a tracker
Write down the exact product surface you were using: ChatGPT web or desktop, Codex, Claude.ai, Claude Code, Grok chat, Grok Bot, or an API console. The same brand name can hide several meters.
Copy the limit message if you can. Wording like “try again later,” “weekly limit,” “rate limit,” or “usage credits” points to different remedies. A 429 from an API console is not the same event as a subscription session cap.
Codex and ChatGPT usage
Open the product’s usage or settings area and look for the plan allowance that applies to the feature you were using. Official pricing and usage documentation describe how paid plans expose usage; exact labels move over time, so trust the live UI over a screenshot from a blog post.
Ask three questions on that screen: which model or feature consumed the quota, whether a reset time is printed, and whether any exceptional credit or banked reset notice appears separately from the ordinary window.
Use the Codex tracker on this site only as a secondary check for public reset or banked-credit announcements. It will not show your remaining messages.
Claude usage
In Claude, start at Settings → Usage. Subscription plans use session windows, and paid plans can also carry weekly limits. Refreshing one window does not automatically clear another.
If you are on an eligible paid plan and the product offers usage credits, treat that as a billed continuation option — not a free manual reset. API rate limits belong in the API console and docs, not in the consumer subscription window.
The Claude tracker here archives exceptional public posts. The Claude limits guide explains the ordinary session and weekly structure with official source links.
Grok chat, Grok Bot and xAI API
For Grok chat, check the in-product plan and limit messaging. Fixed public message caps are not always available as static documentation, so the live product is the authority.
For Grok Bot, read any reset announcement’s stated audience carefully. A Bot-user reset is evidence for Bot, not a blanket reset for every Grok surface.
For the xAI API, open the Console and the rate-limit docs. Throughput limits and HTTP 429 responses are operational constraints, not a social countdown.
When a public tracker helps
A tracker helps after you already understand which meter failed. It can confirm whether an exceptional announcement landed today, preserve the original wording, and show how rare such posts have been in the captured history.
A tracker does not help as a substitute for the usage screen, as a push-notification service, or as a way to force a provider to refill your account. Those jobs belong to the product and its support channels.
Frequently asked questions
I refreshed the page and still cannot send messages. What next?
Another meter may still be exhausted, or the product may be having an outage. Re-check usage, confirm you are on the intended account, and consult the provider status page if the UI looks unhealthy.
Should I upgrade to reset my window?
An upgrade can change future allowances; it does not automatically mean the current exhausted window is wiped. Read the plan terms before paying for that reason alone.
Where do I report a wrong tracker entry?
Use the contact address on the About page and include the original post URL.