Model Context Protocol (mcp) Server + Strava Oauth

This project is an MCP server based on Cloudflare Workers, integrating the Strava OAuth authentication function, allowing users to connect to the MCP service through their Strava accounts.

Health and wellness Developer tools #OAuth authentication #MCP protocol #Cloud service #Sports data .TypeScript

rating : 2.5 points

downloads : 18

update time : 2025-04-23

What is the Strava MCP server?

This is a special AI connection server that allows users to log in with their Strava sports accounts and securely connect their sports data to AI assistants (such as Claude). The server acts as a bridge between Strava and AI applications, ensuring secure and reliable data transmission.

How to use this service?

Simply configure the server address in a supported AI application (such as Claude Desktop), and then connect your Strava account through a simple authorization process. After that, the AI assistant can access your sports data with your permission.

Use cases

It is particularly useful when you want the AI assistant to provide personalized suggestions based on your sports data, such as analyzing running performance, formulating training plans, tracking fitness progress, or generating sports reports.

Main features

Strava account integrationSecurely connect your Strava account using the industry-standard OAuth 2.1 protocol without sharing your password

Data privacy protectionAll sports data is encrypted, and you can revoke access at any time

Real-time data synchronizationKeep data updated in real-time through SSE (Server-Sent Events) technology

Cross-platform supportCompatible with various AI clients such as Claude Desktop and Cursor

Advantages and limitations

Advantages

Add Strava data support to AI applications without development

Globally distributed deployment based on Cloudflare Workers with low latency

Compliant with the latest OAuth 2.1 security standards

Open-source code library, transparent and trustworthy

Limitations

Dependent on the rate limit of the Strava API (200 requests per 15 minutes)

Requires basic server deployment knowledge

Currently only supports the Strava platform

How to use

Visit the Strava developer settings page to create a new application and obtain an API key

Deploy the MCP server

Use the Wrangler CLI tool to deploy the server to Cloudflare Workers

Configure the AI client

Add the server connection information in AI applications such as Claude Desktop

Complete OAuth authorization

Initiate the authorization process in the AI application, log in to your Strava account, and grant permissions

Usage examples

Training analysisLet the AI analyze your recent running data and provide performance evaluation and improvement suggestions

Activity summaryAutomatically generate a monthly sports summary report

Frequently Asked Questions

Will my Strava password be stored on the server?

How to revoke the AI assistant's access to my Strava data?

Will the server store my sports data?

Why can't the AI sometimes get the latest data?

Related resources

Model Context Protocol official website

Learn about the official documentation and specifications of the MCP protocol

Strava developer documentation

The official development documentation of the Strava API

GitHub repository

The source code and issue tracking of this project

Cloudflare Workers documentation

Learn how to deploy and configure Workers

🚀 Model Context Protocol (MCP) Server + Strava OAuth

This is a server based on the Model Context Protocol (MCP), supporting remote MCP connections and featuring built - in Strava OAuth login functionality. Users can connect to your MCP server and log in using their Strava accounts.

🚀 Quick Start

Prerequisites

A Strava account
A Cloudflare account
Node.js and npm installed
Wrangler CLI installed (npm install -g wrangler)

Steps

Clone the repository:

git clone https://github.com/kw510/strava-mcp.git
cd strava-mcp
npm install

Set up your Strava API credentials (see Setting up Strava API Credentials).
Set up a Cloudflare KV namespace:
```
wrangler kv:namespace create "OAUTH_KV"
```
Update the wrangler.toml file with the generated KV ID.
Deploy to Cloudflare:
```
wrangler deploy
```

✨ Features

OAuth Provider

Implements the OAuth 2.0 flow using the workers - oauth library.
Supports Strava's authorization code flow.
Configures callback URLs and client ID/secret for authentication.

Persistent MCP

Stores connection states in Cloudflare KV storage.
Manages persistent sessions and user data using the durable - object library.
Implements secure session management and state storage.

MCP Proxy

Creates an MCP proxy using the @cloudflare/mcp - proxy library.
Supports bidirectional communication via WebSocket and HTTP.
Provides access to remote models and handles requests and responses.

📦 Installation

Setting up Strava API Credentials

Production Environment

Go to Strava's API settings and create a new application.
Configure your application:
- Application Name: Choose a name for your application.
- Category: Select an appropriate category.
- Website: Your website URL.
- Application Description: Briefly describe your application.
- Authorized Callback Domain: mcp - strava - oauth.<your - subdomain>.workers.dev
- Authorized Callback URL: https://mcp - strava - oauth.<your - subdomain>.workers.dev/callback

Set your production environment variables:

wrangler secret put STRAVA_CLIENT_ID
wrangler secret put STRAVA_CLIENT_SECRET

Development Environment

Go to Strava's developer console and create a new application.
Configure your application:
- Application Name: e.g., MCP - Dev
- Description: A short description like MCP Development App
- Website URL: http://localhost:3000
- Authorized Callback URL: http://localhost:3000/login/callback

💻 Usage Examples

Testing Your MCP Server

Using Inspector

Install and start Inspector.
Open Inspector's console.

Create a new MCP connection:

mcp create connection --name test - connection --url http://localhost:5000

Start the connection:

mcp start - connection test - connection

In Inspector, you should see logs and messages from the MCP server.

Using Claude Desktop

Open Claude Desktop and connect to your Cloudflare account.
Navigate to the Compute tab and select your MCP application.

In the right - hand console, you can interact with the MCP proxy:

mcp create connection --name test - connection --url http://localhost:5000
mcp start - connection test - connection

Development

Local Development

Clone the repository and install dependencies.
Start the development server:
```
npm run dev
```
Open your browser and visit http://localhost:5000.

API Rate Limits

Strava's free tier has a limit of 18,000 calls per hour.
Exceeding the limit may result in requests being rejected or delayed.

🔧 Technical Details

Overview

The MCP server (powered by Cloudflare Workers) plays two roles:

As an OAuth service provider for your MCP clients.
As an OAuth client for Strava's OAuth service.

This project serves as a reference example, demonstrating how to integrate an OAuth provider into an MCP server deployed on Cloudflare using the [workers - oauth - provider library](https://github.com/cloudflare/workers - oauth - provider).

📄 Notes

⚠️ Important Note

Ensure that all callback URLs are configured correctly and match exactly the settings in your Strava application.

Regularly check API call limits to avoid exceeding the quota.

Use HTTPS and configure SSL certificates in the production environment to ensure security.

Manage sensitive information (such as client IDs and secrets) using environment variables.

By following the above steps, you can successfully deploy and run an MCP server supporting Strava login, and leverage Cloudflare's powerful features for expansion and optimization.