README - pierre_mcp_server by async-io

Pierre Fitness API

Active Development - This project is under active development and not yet ready for production use. We welcome feedback, contributions, and early testing from the community. Please report issues and share your experiences.

An open source multi-protocol fitness data API providing secure access to fitness data from multiple providers (Strava, Fitbit) through the Model Context Protocol (MCP), A2A (Agent-to-Agent) Protocol, and REST APIs. Built for LLMs and AI applications with comprehensive API key management, rate limiting, OAuth integration, and real-time analytics.

Fitness Data Analysis

Pierre connects fitness data to AI assistants like Claude, ChatGPT, and any agents supporting the MCP/A2A protocol, providing intelligent analysis with location, weather, and performance context.

Analysis Type	Example Queries	Key Features
Running	"What was my longest run this year and where?"	Pace trends, location context, terrain analysis
Cross-Training	"Compare my cycling vs running performance"	Multi-sport analysis, heart rate zones, consistency
Location Intelligence	"Where do I perform best?"	GPS tracking, terrain impact, route optimization
Weather Impact	"How does weather affect my performance?"	Temperature correlation, seasonal patterns
Activity Intelligence	"Analyze my marathon with full context"	Insights with environmental factors
Goal Tracking	"How close am I to my 1000km goal?"	Progress monitoring, achievement analysis
Performance Trends	"Find patterns in my training data"	Long-term analysis, optimization suggestions

See detailed examples: Check out our with 50+ real-world queries.

Integration Methods

Pierre supports multiple integration patterns for different use cases:

Integration Type	Best For	Setup Complexity	Authentication
MCP Protocol (2025-06-18)	AI assistants (Claude, ChatGPT, any MCP compliant client)	Low	JWT Token
A2A Protocol (v0.2.3)	AI agents & applications	Medium	API Keys
REST API	Web apps & dashboards	Medium	OAuth2 + JWT
Single-Tenant	Personal local use	Minimal	Optional

Quick Setup Examples

AI Assistant Integration (Claude, ChatGPT, etc.)

Configure MCP Server

// For Claude Desktop (~/.claude/claude_desktop_config.json)
{
  "mcpServers": {
    "pierre-fitness": {
      "command": "path/to/pierre-mcp-server",
      "args": ["--single-tenant", "--port", "8080"]
    }
  }
}

// For ChatGPT or other MCP-compatible clients
// Use the same MCP protocol with your client's configuration

Connect to Strava
- Visit the OAuth URL provided by Pierre
- Authorize access to your Strava data
- Start asking questions in natural language
Works with any MCP/A2A compatible agent
- Claude Desktop, ChatGPT with MCP support
- Custom AI agents, GitHub Copilot extensions
- Any application supporting MCP or A2A protocols

A2A Integration for Developers

# Register your application
curl -X POST https://your-pierre-server.com/a2a/clients \
  -H "Authorization: Bearer JWT_TOKEN" \
  -d '{"name": "My Fitness App", "description": "AI fitness coach"}'

# Execute fitness tools
curl -X POST https://your-pierre-server.com/a2a/execute \
  -H "Authorization: Bearer API_KEY" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools.execute",
    "params": {
      "tool_name": "get_activities",
      "parameters": {"provider": "strava", "limit": 10}
    }
  }'

Detailed guides: See our for complete integration examples.

Documentation

Guide	Description	Key Topics
	Complete doc guide	All documentation organized by use case
	Setup and configuration	Local setup, OAuth config, Docker deployment
	Complete API documentation	MCP tools, HTTP endpoints, error handling
	Interactive API reference	Complete API spec with examples
	Database setup	SQLite/PostgreSQL setup and database plugins
	Production deployment	Docker, Kubernetes, cloud platforms

Key Features

Category	Features
Integrations	Strava, Fitbit • MCP Protocol • A2A Protocol • REST APIs
Security	OAuth2 + PKCE • JWT Authentication • Encrypted storage • Rate limiting
Intelligence	Activity analysis • Location detection • Weather integration
Architecture	Single & multi-tenant • Cloud-ready • Extensible • Deployment-ready

Quick Start

Method	Command	Use Case
Local	`cargo run --bin pierre-mcp-server -- --single-tenant`	Personal use, development
Docker	`./docker-compose-with-envrc.sh up`	Easy deployment, cloud-ready
AI Assistants	Add to MCP client config	Claude, ChatGPT, agent integration

One-Minute Setup

# 1. Clone and build
git clone https://github.com/Async-IO/pierre_mcp_server.git
cd pierre_mcp_server && cargo build --release

# 2. Run locally (single-tenant mode)
cargo run --bin pierre-mcp-server -- --single-tenant

# 3. Configure AI Assistant (Claude, ChatGPT, etc.)
echo '{
  "mcpServers": {
    "pierre-fitness": {
      "command": "'$(pwd)'/target/release/pierre-mcp-server",
      "args": ["--single-tenant", "--port", "8080"]
    }
  }
}' > ~/.claude/claude_desktop_config.json

Architecture

Mode	Best For	Features
Single-Tenant	Personal use	Local config, no auth required, simple setup
Multi-Tenant	Cloud deployment	JWT auth, encrypted storage, user isolation

License

This project is dual-licensed under either of:

Apache License, Version 2.0 ()
MIT License ()

at your option.

Contributing

We welcome contributions! See for details on how to get started.