claudeusage-mcp

claudeusage-mcp

Provides real-time visibility into Claude Pro and Max subscription usage limits directly within Claude Code by utilizing local OAuth tokens. It enables users to monitor session and weekly usage across different models and receive alerts regarding rate-limiting status.

Category
Visit Server

README

claudeusage-mcp

An MCP server that gives you real-time visibility into your Claude Pro/Max subscription usage — directly inside Claude Code.

No API keys. No scraping. No browser automation. It reads the OAuth token that Claude Code already stores on your machine and calls Anthropic's usage endpoint to get the exact same data shown on claude.ai/settings/usage.


What it shows

Metric Description
Current Session 5-hour burst window utilization with reset countdown
Weekly Usage - All Models 7-day rolling limit across all models
Weekly Usage - Opus Only 7-day Opus-specific usage (Max plans)
Weekly Usage - Sonnet Only 7-day Sonnet-specific usage
Extra Usage Enabled/disabled, monthly limit, credits used

MCP tool output (get_usage)

=== Claude Plan Usage ===

Current Session (5-hour window)
  ████████░░░░░░░░░░░░  37%
  Resets in 2h 14m (Sat Mar 22 07:59PM)

Weekly Usage - All Models (7-day) [HIGH]
  ██████████████████░░  89%
  Resets in 4d 3h (Wed Mar 26 05:00PM)

Weekly Usage - Opus Only
  N/A (not on your plan)

Weekly Usage - Sonnet Only
  ░░░░░░░░░░░░░░░░░░░░  1%
  Resets in 4d 3h (Wed Mar 26 08:59PM)

Extra Usage: Disabled

--- Color Grading ---
0-49% = GREEN (safe)  |  50-79% = ORANGE (moderate)  |  80-99% = RED (high)  |  100% = RATE LIMITED

(live data)

Rate-limited state

When any metric hits 100%, the output changes to show reset date and time:

Weekly Usage - All Models (7-day) [RATE LIMITED]
  XXXXXXXXXXXXXXXXXXXX  100%
  Resets in 1d (Mon Mar 24 05:00PM)

Statusline output

The included statusline script adds a third line to your Claude Code status bar with progress bars and color coding:

  Usage: Session █░░░░░░░░░ 13% | Weekly █████████░ 94% | Sonnet █░░░░░░░░░ 8%

When rate-limited:

  Usage: Session: RATE LIMITED, Resets in: 3h 37m (Sun Mar 22 07:59PM) | Weekly: RATE LIMITED, Resets in: 1d (Mon Mar 23 05:00PM)

Color grading

Range Color Status
0-49% Green Safe
50-79% Orange Moderate
80-99% Red High
100% Red + text RATE LIMITED with reset datetime

Tools

Tool Purpose
get_usage Full dashboard — session, weekly, per-model, extra usage with progress bars
get_session_usage Current 5-hour session window only
get_weekly_limits 7-day limits with per-model breakdown (All, Opus, Sonnet)
check_rate_status Am I about to be rate-limited? Returns LOW / MODERATE / HIGH / RATE LIMITED

Requirements

  • Node.js >= 18
  • Claude Code logged in with a Pro or Max subscription
  • macOS or Linux (reads credentials from Keychain or ~/.claude/.credentials.json)

Installation

Option 1: claude mcp add (recommended)

git clone https://github.com/OrelliusAI/claudeusage-mcp.git
cd claudeusage-mcp
npm install && npm run build

claude mcp add claudeusage -- node $(pwd)/dist/index.js

Restart Claude Code. Then ask "what's my usage?" or type /usage.

Option 2: Manual config

git clone https://github.com/OrelliusAI/claudeusage-mcp.git
cd claudeusage-mcp
npm install && npm run build

Add to your MCP config (.claude.json or project .mcp.json):

{
  "mcpServers": {
    "claudeusage": {
      "command": "node",
      "args": ["/absolute/path/to/claudeusage-mcp/dist/index.js"]
    }
  }
}

Optional: Statusline integration

Add usage data to your Claude Code statusline (always visible at the bottom):

# If you DON'T have a statusline yet:
cp statusline.sh ~/.claude/usage-statusline.sh
chmod +x ~/.claude/usage-statusline.sh

Then set in ~/.claude/settings.json:

{
  "statusLine": {
    "type": "command",
    "command": "bash ~/.claude/usage-statusline.sh"
  }
}

If you already have a statusline, use statusline-combined.sh instead — it pipes your existing statusline and appends the usage line.

How it works

Claude Code (your session)
    |
    |-- Reads OAuth token from:
    |   |-- macOS Keychain ("Claude Code-credentials")
    |   +-- ~/.claude/.credentials.json (Linux/WSL)
    |
    |-- Calls: GET https://api.anthropic.com/api/oauth/usage
    |   Headers: Authorization: Bearer <your-oauth-token>
    |            anthropic-beta: oauth-2025-04-20
    |
    |-- Caches response for 60 seconds
    |   +-- On 429 (rate limit): serves stale cache
    |
    +-- Returns formatted usage data via MCP tools

Zero configuration. The OAuth token is created automatically when you sign into Claude Code with your Pro/Max subscription. No API keys, no environment variables, no .env files.

Authentication

This server uses the same OAuth token that Claude Code creates when you log in:

Platform Location
macOS Keychain (service: Claude Code-credentials)
Linux / WSL ~/.claude/.credentials.json

Required OAuth scopes: user:inference, user:profile

Token issues?

If the server reports authentication errors:

  1. Your token may be expired — restart Claude Code to trigger a fresh OAuth login
  2. Token created via claude setup-token or /login may have wrong scopes — delete credentials and restart Claude Code

Do NOT use API keys (sk-ant-api...). This server is for Pro/Max subscribers, not API users.

Rate limiting

Anthropic's /api/oauth/usage endpoint is aggressively rate-limited. This server handles it:

  1. Caching — responses cached for 60 seconds (MCP) or 120 seconds (statusline)
  2. Stale fallback — on 429, the last successful response is served
  3. Graceful errors — clear messages when the API is unavailable

Who is this for?

Claude Pro and Max subscribers who use Claude Code and want to see their plan usage without opening a browser. This is not for API users — if you have an API key (sk-ant-api...), use the Usage and Cost API instead.

License

MIT

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured