arvo-mcp is a fitness tracking MCP server used to connect AI assistants with Arvo (an AI-powered fitness coach), allowing users to access and manage their training data through natural conversation.
Purpose
This tool is suitable for users who want to track their fitness progress with the help of an AI assistant. It allows Claude Desktop, Cursor, Windsurf, or other MCP-compatible clients to directly retrieve training history, personal records, body composition, and training plans. Typical uses include asking "What should I train today?" and getting a personalized training schedule, or using AI to analyze training progress, monitor PRs, training volume, and muscle group distribution, as well as analyze trends in body composition over time.
Main capabilities
- Provides 29 fitness tools, covering profile, workouts, split plans, personal records (PRs), body progress, and more.
- Supports read and write access: you can both view data and modify data through natural language conversation.
- Supports secure API key authentication, with keys stored locally and never sent to third parties; all communication uses HTTPS.
- Compatible with any MCP client, including Claude Desktop, Cursor, Windsurf, etc.
- 19 read-type tools and 10 write-type tools.
How to use
-
Get an API key: register or log in at arvo.guru, open Settings → API Keys in the web app (if a simplified setup is shown, switch to Advanced mode first, or directly open https://arvo.guru/settings#api-keys），点击 Create Key and copy the key. Note that API key management is currently limited to the web, and is not yet available in the mobile app.
-
Configure the MCP client: add the configuration in claude_desktop_config.json. Configuration file locations: on macOS, ~/Library/Application Support/Claude/claude_desktop_config.json; on Windows, %APPDATA%\Claude\claude_desktop_config.json. You can also add the configuration to Cursor's MCP settings.
-
Start the conversation: after restarting the AI client, you can try asking questions such as: "What's my workout for today?", "Show me my recent PRs", "How's my progress on bench press?", "What muscle groups am I training this week?"
Use cases
- Fitness enthusiasts checking the day's training plan before working out.
- Analyzing long-term training progress, volume, and muscle group distribution with AI.
- Tracking personal records (PRs) and changes in body composition.
- Modifying training plans, preferences, or saving memories through natural language.
Limitations and notes
- An Arvo account is required to generate an API key; it cannot be used without an account.
- It cannot be used offline: arvo-mcp requires an internet connection to communicate with the arvo.guru server; although the MCP server runs locally, data is retrieved from the cloud.
- API keys are stored using SHA-256 hashing, can be revoked at any time from the Arvo dashboard, and support configurable scopes (read-only or read-write).
- How to update: if using npx -y arvo-mcp, the latest version will be fetched automatically; for a global installation, run npm update -g arvo-mcp.
- Security vulnerability reports can be sent by email to security@arvo.guru; see SECURITY.md for details.
The project uses the MIT license. For more information, see arvo.guru, arvo.guru/docs, GitHub Issues, modelcontextprotocol.io, and npmjs.com/package/arvo-mcp.
