> ## Documentation Index
> Fetch the complete documentation index at: https://docs.chrisdoc.dev/hevy-mcp/llms.txt
> Use this file to discover all available pages before exploring further.

# Hevy MCP: AI-powered access to your Hevy fitness data

> Connect Claude, Cursor, Codex, and other MCP clients to your Hevy workouts, routines, and body measurements through 25 ready-to-use tools.

Hevy MCP is an open-source [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for the [Hevy](https://www.hevyapp.com/) fitness tracking app. It lets AI assistants read, analyze, create, and update your workouts, routines, exercise templates, and body measurements — all through authenticated Hevy API requests.

Connect to the **hosted Cloudflare Worker** in seconds with no local install, or run the server locally via `npx`, `bunx`, or Docker. A Hevy PRO subscription and API key are required.

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Connect your first MCP client to Hevy in under five minutes.
  </Card>

  <Card title="Tools Reference" icon="wrench" href="/tools/overview">
    Browse all 25 tools — workouts, routines, exercises, and more.
  </Card>

  <Card title="Connecting Clients" icon="plug" href="/clients/hosted-endpoint">
    Set up Claude Desktop, Codex, Cursor, or any MCP-compatible client.
  </Card>

  <Card title="Guided Prompts" icon="comment-dots" href="/guides/prompts">
    Use built-in prompts for training analysis and workout logging.
  </Card>
</CardGroup>

## How it works

Hevy MCP sits between your AI assistant and the Hevy API. Every tool call is authenticated using your Hevy API key — the server never stores it.

```
Hosted:  AI assistant  →  Streamable HTTP  →  Cloudflare Worker  →  Hevy API
Local:   AI assistant  →  MCP over stdio   →  local hevy-mcp     →  Hevy API
```

**Hosted endpoint** — The production server runs at `https://hevy.chrisdoc.dev/mcp`. Send your Hevy API key as a bearer token on every request. Nothing to install or keep running.

**Local server** — Run `npx -y hevy-mcp` with `HEVY_API_KEY` in the environment. Works with any stdio MCP client including Claude Desktop, Cursor, and Codex.

## Get started in three steps

<Steps>
  <Step title="Get your Hevy API key">
    Create an API key in the Hevy app. API access requires a **Hevy PRO** subscription. Keep the key secure — treat it like a password.
  </Step>

  <Step title="Connect your MCP client">
    Use the hosted endpoint (no install needed) or run locally with `npx`. See [Connecting Clients](/clients/hosted-endpoint) for client-specific setup.
  </Step>

  <Step title="Ask your first question">
    Try: *"Give me a training summary for the last four weeks."* or *"What routines do I have saved on Hevy?"*
  </Step>
</Steps>

## What you can do

<CardGroup cols={2}>
  <Card title="Analyze training progress" icon="chart-line">
    Summarize 1–12 weeks of workouts and body measurements in a single tool call with `get-training-summary`.
  </Card>

  <Card title="Log and update workouts" icon="dumbbell">
    Create completed workouts and update existing ones using `create-workout` and `update-workout`.
  </Card>

  <Card title="Manage routines" icon="list-check">
    Create, update, and search routines and routine folders to keep your training plans organized.
  </Card>

  <Card title="Search the exercise catalog" icon="magnifying-glass">
    Find exercise template IDs by name using `search-exercise-templates` with a 5-minute catalog cache.
  </Card>
</CardGroup>

<Note>
  Hevy MCP requires a **Hevy PRO** subscription for API access. Get your API key from the Hevy app settings before connecting.
</Note>