Turn Your Strava Data Into an AI-Powered Training Assistant

The Strava MCP server transforms your workout data into something your AI assistant can actually understand and work with. Instead of manually digging through Strava's web interface or wrestling with API endpoints, you can now ask Claude, ChatGPT, or any MCP-compatible AI: "Show me the power data from yesterday's ride" or "Find challenging climbs near Boulder" and get instant, detailed responses.

Why This Changes Everything

If you're tired of switching between your AI chat and Strava tabs just to analyze your training data, this server eliminates that friction entirely. Your AI assistant gets direct access to your activities, segments, routes, and detailed performance streams—all through natural conversation.

Before: "Let me check Strava... copy some numbers... paste them into Claude... ask for analysis..."
After: "Analyze my last three rides and create a power profile comparison chart."

What You Get Access To

This isn't just basic activity data. The server exposes nearly everything Strava offers:

• Activity Analysis: Recent workouts, detailed metrics, time-series data streams (heart rate, power, cadence, GPS)
• Performance Tracking: Year-to-date stats, personal records, zone analysis
• Route Management: Saved routes with GPX/TCX export for training platforms
• Segment Intelligence: Starred segments, effort comparisons, geographic segment discovery
• Advanced Data: Lap breakdowns, activity streams with configurable resolution

You can pull power curves from specific intervals, compare segment efforts across months, or export routes for Zwift—all through conversation.

Real Training Scenarios

Power Analysis: "Get the power data from my Tuesday interval session and show me the normalized power for each interval."

Route Planning: "Find steep climbing segments near Flagstaff Road and export my Boulder Loop route as GPX."

Performance Tracking: "Compare my heart rate zones from this month's runs versus last month."

Segment Hunting: "Show me challenging cycling segments between these coordinates with category 3+ climbs."

The server handles Strava's OAuth complexity, token refreshing, and API rate limits automatically. Your AI gets clean, formatted data that it can immediately analyze, visualize, or use for training recommendations.

Developer-Friendly Setup

The authentication flow is streamlined with an included setup script. Create your Strava app, run the auth script, restart Claude—you're connected. The server maintains tokens automatically, so this is a one-time setup.

Configuration requires just your Claude settings file and a few environment variables. The included auth script walks you through the OAuth flow step-by-step.

Integration That Just Works

This server plugs directly into Claude Desktop or any MCP-compatible AI platform. Once configured, your training data becomes part of your AI's knowledge base for every conversation. No switching contexts, no manual data entry.

The natural language interface means you can ask complex questions: "Create a coaching analysis of my last ride with interactive charts showing power, heart rate, and elevation" and get a complete HTML dashboard with professional insights.

Your Strava data transforms from a passive log into an active training partner that your AI can query, analyze, and visualize on demand. Install once, benefit from every conversation.