Usage Monitor for Claude MCP
A Windows and Linux tray app shows Claude rate limits and reports the remaining quota to an agent via a built-in MCP server.
- 332 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The app reads the local Claude Code OAuth token and sends it to `api.anthropic.com` to request limits.
Install
- Works in
Manual install
winget install jens-duttke.usage-monitor-for-claudeThe first README install option for Windows. On Linux the app runs from source.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The app reads the local Claude Code OAuth token and sends it to `api.anthropic.com` to request limits.
This is third-party code. Review the repository files before installing.
What it does
Usage Monitor for Claude watches Claude subscription limits from the system tray. The icon shows two progress bars (session and week), and the popup lists reset countdowns, extra quotas and the installed Claude Code versions. A built-in MCP server gives Claude Code and Claude Desktop a `get_usage` tool. The agent can check the remaining quota itself and warn when heavy work will not fit the limit. Answers come from the app's own data, so no extra requests go to Anthropic. The server stays off until enabled in settings.
Who it is for. Claude Pro, Max and Team subscribers who work in Claude Code daily and do not want to hit a limit in the middle of a task.
Good fit when
- Agents are launched in batches and the remaining quota matters before the start.
- The weekly limit drains faster than time passes and a visible marker helps.
- Claude Code runs under several accounts and each needs its own monitor instance.
- A command must run when a quota resets, such as a phone notification or a fresh five-hour session.
Not a fit when
- Work goes through the Anthropic API with pay-per-token billing, which has no subscription limits.
- Claude Code is not installed or logged in, so the app has no token to use.
- Company policy forbids programs from reading local Claude Code credentials.
Example request
Check the remaining weekly limit through get_usage and tell me whether it covers launching five agents.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
The app reads the OAuth token that Claude Code stores locally and contacts only `api.anthropic.com`. Windows has a ready EXE, while Linux has no binary and runs from source through GTK and WebKit. Some antivirus engines flag the EXE as a false positive from the PyInstaller build, and the author documents signature and hash checks. The MCP server works only while the app itself is running.
How to disable. Quit the app from the tray and turn off `mcp_server` in the settings file. Then remove the MCP entry from Claude Code and delete the EXE. A WinGet install is removed like any package.
FAQ
What does the get_usage tool return?
A list of quotas, each with `used_percent`, `remaining_percent`, a reset time and a `vs_expected` deviation from the time marker. Data comes from the app's last update.
Is an API key needed?
No. The app reuses the Claude Code login and reads the token from its local file.
How do I connect the MCP to Claude Code?
Enable `mcp_server` in settings, restart the app and run `claude mcp add usage-monitor-for-claude --scope user -- "C:\Path\To\UsageMonitorForClaude.exe" --mcp` with your own path.
Does the agent check the quota on its own?
Not every time. Add a matching rule to `CLAUDE.md` so the call happens before heavy work.
What about several accounts?
Run one instance per account with the `--config-dir` parameter and register each MCP under its own name.
Discussion
No questions or reviews yet.
Related
MCP serversA real-time terminal monitor for Claude Code usage: tokens, messages and cost in the five-hour window, with forecasts and export
A limits and token spend monitor for Claude Code, Codex, Antigravity and Grok CLI: macOS menu bar, Windows tray, HTML reports
A Claude Code plugin: a status line with context, 5-hour and 7-day limits, cost and usage of Codex, Gemini and z.ai
A CLI and MCP server that tally token spend across 41 tools and agents, break it down by project and model, and help cut wasted tokens
Claude Resets: usage reset tracker
Claude Resets
A Claude Code usage reset tracker over MCP: the agent says how many days have passed since the last reset and pulls the archive with links to posts
A Claude Code status line showing 5-hour and 7-day quota use plus the pace: whether what is left lasts until the window resets
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Usage Monitor for Claude (this card) | no | not required | 332 | |
| ccusage | no | 19,337 | counts token spend from local logs; Usage Monitor shows live remaining subscription limits and answers the agent through MCP | |
| Claude Code Usage Monitor | no | 8,792 | tracks token consumption in the terminal; this is a tray app with an MCP server for the agent |
README in short
The app runs without installation and without manual token setup. Alerts, event commands and the usage forecast are configured separately.