A practical guide to Claude and Codex usage limits
Read session, weekly and additional allowances, understand saved readings, and plan a reset reminder for Claude or Codex.
Updated September 6, 2026
A useful usage check answers three questions: which limit applies, how recent is the reading, and when should I check again? This guide focuses on the two providers supported by How Much Usage: Claude and Codex.
Read the full account card
- Session and longer windows: review every reported allowance, not just the first bar.
- Additional limits: model or feature restrictions can matter even when another bar has room.
- Reading age: saved usage is a snapshot. Work in other sessions can change the account after the last check.
- Reset time: a countdown follows a provider-reported time; it is not a new usage reading.
Use the provider’s own view when needed
Claude offers Usage settings, and Codex offers its own usage dashboard and CLI status view. Our Claude guide and Codex guide link to current official documentation.
Build a simple checking routine
- Press Refresh all before choosing where to start.
- Use the account suggestion as a starting point, then inspect any additional limits.
- Download a calendar reminder if a useful reset is known.
- Refresh again after the reset before assuming the allowance is available.
How Much Usage does not collect readings in the background, send push alerts, or cover every AI service. Its compact widget displays saved readings from the dashboard.
Frequently asked questions
- Which providers does How Much Usage support?
- Claude and Codex. Codex support covers reported allowances on connected ChatGPT accounts, not general ChatGPT or OpenAI API usage.
- Can I request another provider?
- Yes. Share which provider and usage information you need through the feedback link. Future support is not guaranteed.