# transition-mcp [Health: Active]

**Category:** 🏃 Sports  
**Repository:** https://github.com/nftechie/transition-mcp  
**GitHub Stars:** 0  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/transition-mcp

## Description
AI-powered multisport coaching — workouts, training plans, and performance analytics.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "transition-mcp": {
    "url": "https://api.transition.fun/api/v1/wod?sport=run&duration=45"
  }
}
```

## Documentation

## What transition-mcp MCP server does

The transition-mcp MCP server connects MCP-compatible clients with Transition, an endurance-training platform for runners, cyclists, swimmers, and triathletes. It exposes training information and coaching functions through MCP rather than requiring an agent to implement the Transition API directly.

The available tools cover both read and action-oriented workflows:

- Retrieve today’s scheduled workout.
- Retrieve a seven-day training plan.
- Request adaptation of the current plan.
- Read CTL, ATL, and TSB fitness metrics.
- Send questions to the AI endurance coach.
- Generate a workout without authentication.

It also exposes athlete information as MCP resources, including the athlete profile, recent activities from the previous 14 days, and performance data such as FTP, thresholds, and PMC information.

## How it works

The transition-mcp MCP server is built as a Go executable in the repository’s `mcp` directory. An MCP client starts that executable over standard input and output, and the server wraps requests to the Transition API as MCP tools and resources.

Transition API requests use the `TRANSITION_API_KEY` environment variable for authenticated features. The generated workout tool is available without an API key. The same underlying service can also be used directly through HTTP, but that direct API mode is separate from running the MCP executable.

The repository additionally provides a Claude Code and OpenClaw skill that calls the API directly over HTTP. That option does not use the MCP protocol or the compiled server, so choose it when the client supports the provided skill workflow instead of MCP configuration.

## Setup and configuration

Build the server from the repository’s `mcp` directory:

```bash
cd mcp
go build -o transition-mcp .
```

Configure an MCP client to launch the resulting binary and provide the API key through its environment settings. In Claude Desktop, the configuration uses an `mcpServers` entry with the executable’s absolute path and a `TRANSITION_API_KEY` value. Restart the client after adding the server configuration.

To obtain a key, install the Transition mobile app, create an account, complete onboarding, then open Settings > API Keys and generate a key. The key is displayed only once. The free account tier allows 100 read requests per day and three AI requests per day; the paid limits are 10,000 reads and 100 AI requests per day.

## Tools and capabilities

The transition-mcp MCP server provides six tools: `get_todays_workout`, `get_week_plan`, `adapt_plan`, `get_fitness_metrics`, `chat_with_coach`, and `generate_workout`. The first five use personalized Transition data or coaching functions. `generate_workout` provides a free Workout of the Day and does not require authentication.

Three resources provide read access to `transition://athlete/profile`, `transition://athlete/recent-activities`, and `transition://athlete/performance`. These resources cover goals and preferences, recent training history, and performance indicators.

## Limitations and notes

Personalized features depend on a Transition account and valid API key. API quotas apply separately to read and AI operations, so automated clients should account for the documented daily limits. The repository is licensed under MIT.

_Full upstream README: https://allmcps.com/mcp/transition-mcp/readme_