Strava MCP Server
# Strava MCP Server
[](https://github.com/yorrickjansen/strava-mcp/actions/workflows/ci.yml)
[](https://codecov.io/gh/yorrickjansen/strava-mcp)
A Model Context Protocol (MCP) server for interacting with the Strava API.
<a href="https://glama.ai/mcp/servers/@yorrickjansen/strava-mcp">
<img width="380" height="200" src="https://glama.ai/mcp/servers/@yorrickjansen/strava-mcp/badge" alt="Strava Server MCP server" />
</a>
## User Guide
### Installation
You can easily install Strava MCP with `uvx`:
```bash
uvx strava-mcp
```
### Setting Up Strava Credentials
1. **Create a Strava API Application**:
- Go to [https://www.strava.com/settings/api](https://www.strava.com/settings/api)
- Create a new application to obtain your Client ID and Client Secret
- For "Authorization Callback Domain", enter `localhost`
2. **Configure Your Credentials**:
Create a credentials file (e.g., `~/.ssh/strava.sh`):
```bash
export STRAVA_CLIENT_ID=your_client_id
export STRAVA_CLIENT_SECRET=your_client_secret
```
3. **Configure Claude Desktop**:
Add the following to your Claude configuration (`/Users/<username>/Library/Application Support/Claude/claude_desktop_config.json`):
```json
"strava": {
"command": "bash",
"args": [
"-c",
"source ~/.ssh/strava.sh && uvx strava-mcp"
]
}
```
### Authentication
The first time you use the Strava MCP tools:
1. An authentication flow will automatically start
2. Your browser will open to the Strava authorization page
3. After authorizing, you'll be redirected back to a local page
4. Your refresh token will be saved automatically for future use
### Available Tools
#### Get User Activities
Retrieves activities for the authenticated user.
**Parameters:**
- `before` (optional): Epoch timestamp for filtering
- `after` (optional): Epoch timestamp for filtering
- `page` (optional): Page number (default: 1)
- `per_page` (optional): Number of items per page (default: 30)
#### Get Activity
Gets detailed information about a specific activity.
**Parameters:**
- `activity_id`: The ID of the activity
- `include_all_efforts` (optional): Include segment efforts (default: false)
#### Get Activity Segments
Retrieves segments from a specific activity.
**Parameters:**
- `activity_id`: The ID of the activity
#### Get Segment Leaderboard
Gets the leaderboard for a specific segment.
**Parameters:**
- `segment_id`: The ID of the segment
- Various optional filters (gender, age group, etc.)
## Developer Guide
### Project Setup
1. Clone the repository:
```bash
git clone <repository-url>
cd strava
```
2. Install dependencies:
```bash
uv install
```
3. Set up environment variables:
```bash
export STRAVA_CLIENT_ID=your_client_id
export STRAVA_CLIENT_SECRET=your_client_secret
```
Alternatively, create a `.env` file with these variables.
### Running in Development Mode
Run the server with MCP CLI:
```bash
mcp dev strava_mcp/main.py
```
### Manual Authentication
You can get a refresh token manually by running:
```bash
python get_token.py
```
### Project Structure
- `strava_mcp/`: Main package directory
- `__init__.py`: Package initialization
- `config.py`: Configuration settings using pydantic-settings
- `models.py`: Pydantic models for Strava API entities
- `api.py`: Low-level API client for Strava
- `auth.py`: Strava OAuth authentication implementation
- `oauth_server.py`: Standalone OAuth server implementation
- `service.py`: Service layer for business logic
- `server.py`: MCP server implementation
- `tests/`: Unit tests
- `strava_mcp/main.py`: Main entry point to run the server
- `get_token.py`: Utility script to get a refresh token manually
### Running Tests
```bash
pytest
```
### Publishing to PyPI
#### Building the package
```bash
# Build both sdist and wheel
uv build
```
#### Publishing to PyPI
```bash
# Publish to Test PyPI first
uv publish --index testpypi
# Publish to PyPI
uv publish
```
## License
[MIT License](LICENSE)
## Acknowledgements
- [Strava API](https://developers.strava.com/)
- [Model Context Protocol (MCP)](https://modelcontextprotocol.io/)TDQS
Scored across 3 tools
Each tool has a clearly distinct purpose with no overlap: get_activity retrieves details of a specific activity, get_activity_segments focuses on segments within an activity, and get_user_activities lists the user's activities. The descriptions clearly differentiate these functions, making tool selection unambiguous for an agent.
All tool names follow a consistent verb_noun pattern with snake_case (get_activity, get_activity_segments, get_user_activities). The naming is predictable and readable, using 'get' as the verb throughout, which aligns well with the read-only nature of these tools.
With only 3 tools, this server feels under-scoped for a Strava integration, which typically involves activities, segments, athletes, and more. While the tools cover some read operations, the count is too low to support comprehensive agent workflows, such as creating or updating activities, which are common in fitness tracking domains.
The tool surface is severely incomplete for a Strava server, as it only includes read operations (get) with no support for create, update, or delete actions. There are significant gaps, such as missing tools for managing segments, athletes, or uploading activities, which will likely cause agent failures when attempting full interactions with the Strava API.