Quick answer
The local whoburnedmore CLI command is npx whoburnedmore --local. Decline its optional publish offer to keep the report local. The standard command signs in and submits daily aggregates. Neither mode submits prompts, source code, or filenames.
What this guide answers
The reader hit a limit or saw a warning and wants a plain-English explanation.
Best for
Claude Code users on Pro and Max who need to plan heavy coding sessions around the five-hour session limit.
Key facts, checked November 8, 2026
- Paid Claude plans have two limits: a session limit that resets every five hours and weekly limits. Both are shared across Claude and Claude Code, including IDE usage. source
- The 5-hour clock starts at your first message, and the limit resets in full five hours later. source
- If you hit the five-hour limit mid-response, Claude Code may finish the current step and show "Usage limit reached · wrapping up". On Pro this applies up to once per weekly period, and on Max each time you reach a five-hour limit. It is rolling out gradually. source
- The 5-hour limit was not affected by the May-September 2026 Claude Code weekly-limit promotion. source
A five-hour session, not a daily cap
The session period starts when you send your first message, and your limit resets in full five hours later. Usage does not drip back hour by hour, and it does not wait for midnight.
- The clock starts at your first message.
- The limit resets in full five hours later.
- Small prompts can still matter inside a tight window.
What drains the window fastest
Large context, repeated tool loops, broad refactors, and pasted logs can burn through a window quickly. The issue is usually not one prompt but an accumulating session.
- Avoid dumping unrelated files.
- Ask for narrower diffs.
- Start a fresh session when the context is stale.
How to work near the limit
Near a limit, switch from open-ended generation to precise review or planning. Shorter tasks reduce the chance that the next request becomes the one that blocks you.
- Use smaller file scopes.
- Run tests yourself before asking the model.
- Save large rewrites for a recovered window.
How to measure the window indirectly
Even if a tool does not expose every limit detail, token burn over time gives you a practical warning system. A visible burn curve is enough to change behavior before lockout.
- Track hourly burn.
- Mark sessions that hit limits.
- Compare burn rate before and after workflow changes.
Read the reset time instead of guessing
Because the session resets in full at a fixed point, the useful number is the reset time. Claude Code gives it to you in three places.
- /usage in the CLI shows plan usage bars.
- A status line script can read rate_limits.five_hour.used_percentage and rate_limits.five_hour.resets_at (Unix epoch seconds) on Pro, Max and Ultra.
- Settings > Usage on claude.ai shows Current session usage and the time remaining.
What to do when you hit it
The message "You've hit your session limit" means a plan-wide window. Switching models with /model will not restore access. A model-specific message such as "You've hit your Opus limit" is different: switching to another model family does keep you working. Your options are to wait for the reset, to turn on usage credits so work continues at usage-credit rates (/usage-credits opens the setting), or, on v2.1.234 or later, to let Claude Code wait and automatically continue the interrupted task after the reset. Admins can control that last behavior with autoContinueAtUsageLimit.
Make a session last the full five hours
Long conversations are the usual cause of an early lockout, because every request resends the conversation. Use /clear between unrelated tasks. Check the /usage breakdown for anything taking 10% or more of your usage, such as a large CLAUDE.md or a busy MCP server. Send routine subagent work to a smaller model. A message after a long break also misses the prompt cache and reprocesses your whole context, so compact or clear before stepping away.
Common questions
Is the Claude Code 5-hour limit rolling?
Not in the drip-back sense. Anthropic says the period starts with your first message and the limit resets in full five hours later.
Does claude.ai chat use the same 5-hour limit as Claude Code?
Yes. On Pro and Max, activity in Claude and Claude Code counts against the same limits.
Which plan has no 5-hour limit?
Ultra, at $200 per month, has no five-hour limit for verified accounts. Weekly limits still apply.
Can whoburnedmore tell me when my window resets?
No. It reads local token totals, not plan state. Use /usage or the status line rate_limits fields for reset times.