Is Claude down, or have I hit a usage limit?

A usage limit is specific to your account and names itself in the error — you will see wording about usage, quota or limits, and it arrives instantly. An outage affects everyone, shows as a connection error, timeout or 5xx server error, and appears on Anthropic's status feed. If isclaudeup.com says Claude is up and you are still blocked, it is almost certainly your own limit.

When do Claude usage limits reset?

Claude subscription plans use rolling windows rather than a fixed daily clock. A short session window frees up a few hours after the messages that consumed it, and a longer weekly window refills on a rolling basis. Because the windows roll, capacity returns gradually rather than all at once at midnight.

Why does Claude Code use up my limit so fast?

Agentic work re-sends context on every step, so a single task that reads twenty files and iterates ten times costs far more than twenty chat messages. Large repositories, long sessions without compaction, and retrying a failing run are the three biggest consumers.

Claude usage limits: why you hit them, and when they reset

The short version: if this site says Claude is up and you are still blocked, you have hit your own limit, not an outage.

A large share of “Claude is down” moments are not outages at all. They are a usage limit with a vague error message, and at 11pm the two feel identical. This page is how to tell them apart in about ten seconds, and what your options are once you know.

Step one: is it you, or is it them?

Check the live status first — it reads Anthropic's official feed. Then match your symptom against this:

It is a usage limit if…

The error names it — anything mentioning usage, limit, quota or 429.
It arrives instantly, with no hanging or spinner. Limiters reject before any work happens.
It is consistent — every request fails the same way, immediately, not one in five.
A different account, or a colleague on the same network, works fine.

It is an outage if…

You see connection errors, timeouts, or 500 / 502 / 503 server errors.
Requests hang before failing, rather than bouncing straight back.
It is intermittent — some requests succeed, some do not.
Other people are complaining at the same moment, and the status feed agrees.

There is a third case that fools almost everyone: degraded but not down. Requests succeed, just slowly, and some time out client-side and look like failures. That is why this site publishes a KINDA verdict rather than forcing everything into a binary.

The limits you can actually hit

Session windows. Subscription plans meter usage across a rolling window of a few hours. This is the one most people mean when they say they have been limited.
Weekly windows. A longer rolling budget sitting on top of the session window. Heavy agentic users hit this one without ever hitting the session cap.
Per-minute request and token limits. Burst limits that matter for scripts and parallel agents, and almost never when you are typing.
Context window. Not a rate limit, but it surfaces as a failure part-way through a long session. If it always breaks at roughly the same point, suspect this rather than a quota.

Exact figures vary by plan and change over time, so we deliberately do not publish numbers that would go stale. Anthropic's rate limit documentation and your own account usage page are the only sources worth trusting for current values.

Why Claude Code burns through it faster than chat

This surprises people who move from the chat interface to the CLI. Agentic work re-sends accumulated context on every step. A task that reads twenty files and iterates ten times is not twenty messages of cost — it is closer to two hundred. The three biggest consumers, in order:

Unscoped exploration. Letting the agent search a large repository to find something you could have named directly.
Long sessions without compaction. Every turn carries the whole history forward.
Retrying a failing run. Work already done still counts. Three attempts at a task that dies at 80% costs more than two clean completions.

When does it reset?

Rolling windows, not a midnight clock. Capacity comes back in the same shape you spent it: burn a week's budget in an afternoon and you get it back in one lump a week later; spread it evenly and it trickles back continuously.

Practical consequence: after hitting a limit, wait a while and try one small request rather than hammering retry. If the small one succeeds, some of your window has refilled.

What to do while you wait

If it is an outage, leave your email on the status page and we will send one message the moment Claude recovers — no need to keep refreshing. If it is a usage limit, nothing external will change; the window refills on its own schedule.

Either way it is a good moment for the work that does not need an agent: review the diff you already have, write the test you skipped, or step away. There is also a free browser arcade if you would rather not think about it for ten minutes.

Not affiliated with, endorsed by, or sponsored by Anthropic. Claude and Anthropic are trademarks of Anthropic. Nothing here is official guidance — always check Anthropic's documentation for current limits.