Guide · checked 2 November 2026

Claude Code hit its usage limit mid-task? Here is what to do next

Independent project. Not affiliated with Anthropic or OpenAI.

You were in the middle of a refactor, Claude Code stopped, and the message says you have hit a limit. Your files are fine. Your place in the task is the thing at risk. This page covers what the message means, how to save your place, and your options for getting going again.

The short version

Check which limit you hit and when it resets (/usage).
Save your place yourself: look at git status, commit or stash your work in progress, and write a few lines on what was done and what is next.
Pick a way forward: wait for the reset, add usage credits, upgrade, or carry on in a different coding agent.

Where Limit Rescue fits: step 2 is much easier before the limit hits, because once Claude Code is blocked you can no longer ask it to summarize its own work. Limit Rescue is a small local command line tool that warns you at 80% and 95% of your usage limit and writes a "Where I left off" note while you can still use the agent. Then limitrescue resume codex carries on in Codex. It works with Claude Code and Codex, and it is in early access, not public yet.

Join the list

Limit Rescue v0.1.0 is not on npm yet, has only been tested on macOS so far, and the early-access list is on limitrescue.com. Details are in the section further down.

What the message means

Claude Code shows one of a few messages when a plan quota runs out. Per the Claude Code error reference, these look like "You've hit your session limit", "You've hit your weekly limit", "You've hit your Opus limit" or "You've hit your Sonnet limit", each with a reset time. (Claude Code error reference, Usage limits)

What that page says about them:

Claude Code blocks further requests until the reset time shown in the message.
The session and weekly limits are shared across all models, so switching models with /model does not restore access.
The Opus and Sonnet limits each apply only to that model family, so switching to a model outside the family keeps you working.
Usage counts against the session and weekly allowances at the same time, so a burst of heavy activity can use up the weekly allowance before the session window resets.

On Pro and Max plans, Claude's help center says there is a five-hour session limit and a weekly limit, and that these limits are shared across Claude and Claude Code, so chat in the Claude app counts too. (Use Claude Code with your Pro, Max, or Ultra plan)

The "wrapping up" allowance

The same help center article describes a small wrap-up allowance. If you reach the five-hour limit while Claude Code is partway through a response, it may keep working briefly to finish the step in progress, and you will see "Usage limit reached · wrapping up". It applies only to a response already in progress, it counts toward your weekly limit, and the article says it is rolling out gradually, so you may not have it yet.

Step 1: Find out how long you are stuck

Run /usage to see your plan limits and when they reset. The error reference lists it as the way to check. A session limit that resets in 40 minutes is a very different problem from a weekly limit that resets on Monday.

Step 2: Save your place

Nothing on the Claude or Anthropic pages tells you to do this, so treat it as plain developer habit. It takes two minutes and it is what makes every option below painless.

See what changed. Run git status and git diff --stat.
Protect the work. Commit to a work-in-progress branch, or git stash if you prefer. Do not leave a half-edited tree you cannot describe tomorrow.
Write a short note. Put four things in a file such as HANDOFF.md in the repo: - the goal, in a sentence or two - what is done - what is left, and the very next step - decisions you made and things you tried that did not work
Note the session. Run /rename on the session (for example /rename auth-refactor) if you have not, so you can find it later with claude --resume <name>. The sessions docs cover naming.

If you have no wrap-up allowance, Claude Code will not answer requests after the block, so the note has to come from you and from the git diff.

Step 3: Choose how to keep going

Option A: Wait, and let Claude Code continue by itself

Claude Code's interactive mode docs say that when a claude.ai usage limit stops Claude mid-task, Claude Code waits in the open session and continues the task on its own after the limit resets. It is on by default in interactive sessions signed in with a claude.ai subscription and needs v2.1.234 or later. You can cancel a wait with Esc and turn the feature off. (Wait for a usage limit to reset) Keep the session open for this to work.

Taskhold's guide on the same topic says background sessions, -p runs and weekly resets days away do not get that wait, so you need to resume those yourself. That is Taskhold's reading, so check the Claude Code docs for your setup. (Taskhold: Continue Claude Code after a usage limit)

In the Desktop app, the error reference says the session-limit card has an "Auto-continue when limits reset" checkbox and the weekly-limit card does not.

Option B: Add usage credits or upgrade

The error reference points to /usage-credits to buy additional usage on Pro, Max, and Ultra, or to request it from an admin on Team and Enterprise. The help center article also lists upgrading to a bigger plan as a choice. Check the linked pages for how billing works before you turn it on, since that is a spending decision.

Option C: Resume later in the same session

When the limit resets, the sessions docs say claude --continue reopens the most recent session in the current directory, claude --resume opens the session picker, and claude --resume <name> resumes a named session directly. (Manage sessions) Your HANDOFF.md is still worth having, because a resumed conversation can be long and a short note keeps you oriented.

Option D: Carry on in a different coding agent

If you also have Codex, you can keep the task moving while Claude Code waits for its reset. Neither tool can open the other's conversation, so what you move is context: the git state and your note. The step by step version is in How to switch from Claude Code to Codex mid-task without losing your place.

Two things to know before you do:

Codex has its own usage limits that depend on your ChatGPT plan, and starting a new Codex chat does not reset them (see Using Codex with your ChatGPT plan).
OpenAI says ChatGPT and Codex will share one usage allowance starting in December. We wrote up what that page does and does not say in ChatGPT and Codex are moving to one shared limit.

How to see the limit coming next time

Claude Code can warn you before you run out. The error reference mentions a message such as "You've used 85% of your session limit", and says you can add the rate_limits fields to a custom status line to watch your remaining allowance, or use the usage ring in the Desktop app. Those are worth turning on.

Quick answers

What does "You've hit your session limit" mean in Claude Code?

Claude Code blocks further requests until the reset time shown in the message. The session and weekly limits are shared across all models, so switching models with /model does not restore access. The Opus and Sonnet limits apply only to that model family.

Will Claude Code carry on by itself after the limit resets?

In an interactive session signed in with a claude.ai subscription, Claude Code can wait in the open session and continue the task after the reset. This needs v2.1.234 or later and an open session. You can cancel the wait with Esc.

How do I see how long I am stuck?

Run /usage to see your plan limits and when they reset.

Can Limit Rescue save my place before the limit hits?

That is what it is for. Limit Rescue v0.1.0 warns at 80% and 95% of your usage limit and writes a "Where I left off" note, and limitrescue resume codex continues the work in Codex. It is a local command line tool that makes no network calls, works with Claude Code and Codex, and has only been tested on macOS so far.

Where Limit Rescue helps

Limit Rescue is about the moment between "I am getting close" and "I am blocked", and about what you do after.

A warning at 80% and again at 95% of your usage limit, while you can still ask the agent to tidy up.
A "Where I left off" note, written so another agent can read it cold and carry on.
limitrescue resume codex (or limitrescue resume claude the other way) continues the work in the other agent.
Local only. It is a local command line tool and makes no network calls. Claude Code and Codex still talk to their own servers as they normally do.
Works with Claude Code and Codex.

The goal is that the limit stops being a wall and becomes a handoff.

What it does not do yet

We would rather you hear this from us:

It is v0.1.0 in early access. It is not public yet and not on npm. You join the list on limitrescue.com and we send you the install when more spots open.
It has been tested on macOS only so far.
It supports Claude Code and Codex only, not other agents.
It cannot move a live conversation from one agent to another. Nothing can. It hands over a note and your files.

If you want a side by side look at other tools in this space, see Claude Code and Codex limit tools compared.