You can check your Claude Code allowance frhttps://www.usagebuddy.com/admin/blogom Claude → Settings → Usage. The page shows how much included usage remains and when each limit resets. If you use Claude Code through an API key, check /cost inside Claude Code and the Claude Console instead, because API usage is billed separately from a Claude subscription.

Check your Claude Code usage from the Claude website

For Claude Pro, Max, Team, or Enterprise accounts connected through Claude login:

  1. Open Claude Settings → Usage.
  2. Sign in to the same Claude account you use with Claude Code.
  3. Find the current usage meters and reset times.
  4. If you have more than one account, confirm the account email before trusting the numbers.

The exact meters depend on your plan. A Claude Pro account normally shows a session-based limit and a weekly limit. Anthropic says the session limit resets every five hours, while the weekly limit resets at a fixed day and time assigned to the account. The Usage page shows the next reset time for your account.

Do not estimate the reset from when you started working. Read the time displayed on the Usage page. It is the useful number when you are deciding whether to start a long task now or wait for the allowance to return.

Plan usage and API usage are different

Claude Code can run through a Claude subscription or through an API key. The correct place to check usage depends on how you signed in.

How Claude Code is connected Where to check What you are measuring
Claude Pro, Max, Team, or Enterprise login Claude Settings → Usage Included plan allowance and reset times
Anthropic Console API key /cost in Claude Code and the Console Token-based API spending
Bedrock, Vertex AI, or Microsoft Foundry Provider billing and usage tools Usage billed by that provider

The /cost command is useful for API billing. It reports token and dollar usage for the current session. It is not a replacement for the plan meter shown in Claude Settings.

What the Claude usage limits mean

Claude usage is not a fixed number of prompts. A short question about one file uses less capacity than a long debugging session across a large repository.

Each turn can include:

  • The conversation history from earlier turns
  • Instructions from CLAUDE.md
  • Files Claude has read
  • Tool results and terminal output
  • Your new prompt

The conversation becomes heavier as the session grows. A later message may consume more capacity than an early message even if both prompts are one sentence long.

Model choice matters too. Anthropic recommends Sonnet for most coding work and reserves Opus for harder debugging, architecture, or wide refactors. Opus uses more of the available quota, so leaving it selected for routine edits can shorten the amount of work you get from a usage window.

Image placeholder
Add a screenshot of Claude Settings → Usage showing the session meter, weekly meter, and reset times. Blur the account email and any billing information.
Suggested alt text: Claude Code usage limits and reset times in Claude Settings

Usage limit and context limit are not the same

A usage limit controls how much Claude you can use during a period. A context limit controls how much information Claude can hold in one conversation.

If your plan usage is exhausted, Claude tells you when it will reset or offers the options available for your account. If the conversation context is full, start a new conversation or reduce the current history.

These Claude Code commands help with context:

  • /context shows what is loaded into the current context.
  • /compact summarizes the conversation so you can continue the same task with less history.
  • /clear starts a fresh conversation while keeping the project files and CLAUDE.md available.
  • /model shows the models available to the current account and lets you switch models.

Use /clear between unrelated tasks. Use /compact when the task is still in progress and you need to keep the important history. Copy anything you may need before running /clear, because the command cannot restore the removed conversation.

Why Claude Code usage can drop quickly

The usual cause is a large amount of context being sent repeatedly. Common examples include:

  • Keeping one conversation open across several unrelated tasks
  • Pasting full logs when only the error and nearby lines are relevant
  • Loading large files that Claude does not need
  • Using Opus for mechanical edits
  • Asking Claude to generate a large change before agreeing on a plan
  • Keeping old or overly detailed instructions in CLAUDE.md

A few changes can make the allowance last longer:

  1. Start a fresh conversation when you switch tasks.
  2. Point Claude to a file and function instead of pasting the whole file.
  3. Trim logs to the failing command, error, and relevant stack frames.
  4. Use Sonnet for normal implementation work and switch models only when the task needs it.
  5. Ask for a file-by-file plan before a change that touches several parts of the codebase.

These steps also make the conversation easier to inspect when Claude takes a wrong turn.

What to do when you reach the limit

First, open the Usage page and identify which limit is exhausted. Then choose from the options shown for your plan:

  • Wait for the displayed reset time.
  • Switch to a lighter model if the account still allows it.
  • Use purchased usage credits if they are enabled for your paid plan.
  • Move the task to another account that has available capacity.
  • Use an API key when separate pay-as-you-go billing makes sense for the work.

Do not confuse a full context window with an exhausted account limit. /compact or /clear helps with context. It does not refill the plan allowance.

Checking several Claude accounts without signing in and out

Claude’s Usage page works well for one account. With several personal, client, or team accounts, the manual routine becomes repetitive: switch account, open Usage, remember two percentages and two reset times, then repeat.

Usage Buddy keeps supported Claude Code accounts in one Windows dashboard. Each account can have a clear name, so you can compare its remaining allowance and reset time before choosing where to run the next task. You can also set a warning threshold and receive an alert when an account is running low or ready again.

Usage Buddy also supports Codex and Cursor accounts. Credentials and usage data stay on the computer, and the licence is a one-time $4.99 purchase. If you use one Claude account, the native Usage page may be enough. If you rotate between several AI coding accounts, the combined view removes the repeated account checking.

Frequently asked questions

Where can I see my Claude Code reset time?

Open Claude Settings → Usage while signed in to the account used by Claude Code. The page displays the next reset time for the limits attached to that account.

Does Claude Code have a fixed number of prompts?

No. Consumption depends on the model, conversation length, files, tool results, prompt size, and task complexity. Two prompts can use different amounts of the allowance.

Does /cost show my Claude Pro or Max allowance?

/cost is intended for token and dollar usage under API billing. Check Claude Settings → Usage for the included allowance and reset times on a Claude subscription.

Does /clear reset Claude Code usage?

No. /clear removes the current conversation history. It can reduce the context carried into later turns, but it does not refill an exhausted plan limit.