> ## 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: Connect AI assistants to your Hevy fitness data

> Hevy MCP is an open-source MCP server for the Hevy fitness app. Give Claude, Cursor, and Codex read and write access to your workouts, routines, and body measurements.

Hevy MCP is an open-source [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for the [Hevy](https://www.hevyapp.com/) fitness and workout tracking app. It lets AI assistants read, analyze, create, and update your Hevy workouts, routines, exercise templates, and body measurements through authenticated Hevy API requests — all without leaving your AI client of choice.

Instead of manually exporting CSVs or copying data between apps, you ask your assistant in plain language: *"Summarize my last six weeks of training"* or *"Create a workout from my push-day routine"* — and it does the work by calling Hevy's API on your behalf.

## What is MCP?

The **Model Context Protocol** is an open standard that lets AI assistants connect to external data sources and tools through a well-defined interface. An MCP server exposes a set of typed tools that the assistant can call; the assistant decides when and how to call them based on your instructions.

For fitness data, MCP matters because your workout history lives outside the model's training data. By bridging Hevy's API with MCP, your assistant gains live access to your real training logs, body measurements, and routines — and can reason over them, spot trends, and take action — without you having to copy-paste anything.

## What Hevy MCP can do

<CardGroup cols={2}>
  <Card title="Analyze training progress" icon="chart-line">
    Summarize 1–12 weeks of workout activity and body-measurement trends in a single tool call. Surface consistency gaps, frequently trained exercises, and volume changes.
  </Card>

  <Card title="Log and update workouts" icon="dumbbell">
    Create completed workouts and update existing ones by ID. The assistant collects actual set results before writing — it never invents data.
  </Card>

  <Card title="Manage routines & folders" icon="list-check">
    List, search, create, and update workout routines and routine folders. Load a full routine with its exercise configuration in one call.
  </Card>

  <Card title="Search exercise templates" icon="magnifying-glass">
    Search the full Hevy exercise catalog by title substring. Results are compact and AI-friendly, so discovery does not produce huge responses.
  </Card>
</CardGroup>

## Architecture

Hevy MCP ships in two modes — a hosted Cloudflare endpoint and a local stdio process — but both expose the same 25 tools:

```text theme={null}
Hosted:  Your AI assistant  →  Streamable HTTP  →  Cloudflare Worker  →  Hevy API
Local:   Your AI assistant  →  MCP over stdio   →  local hevy-mcp     →  Hevy API
```

**Hosted endpoint** (`https://hevy.chrisdoc.dev/mcp`) — A production Cloudflare Worker that runs remotely. Your client sends your Hevy API key as a bearer token on every request. The Worker validates the key with Hevy, creates a fresh MCP server per request, and forwards no data outside the request lifecycle. Nothing to install or keep running.

**Local stdio** (`npx -y hevy-mcp`) — The same server runs as a child process on your machine. Your MCP client launches it and communicates over stdin/stdout. The API key is provided through the child-process environment, never in a URL or CLI argument. Suitable for clients that cannot attach a fixed authorization header to remote requests, or when you prefer local control.

## Prerequisites

<Note>
  A **Hevy PRO** subscription is required. API access is gated behind Hevy PRO, and your API key is only available from within the Hevy app to PRO subscribers. There is no free-tier API key.
</Note>

You will also need:

* A **Hevy account** with PRO and an active API key generated in the app
* An **MCP-compatible client** — Claude Desktop, Cursor, Codex, or any client that supports stdio or Streamable HTTP MCP servers
* **Node.js 20 or newer** if you plan to run the local stdio server via `npx`

Ready to connect? Head over to [Quickstart](/hevy-mcp/quickstart) to be up and running in under five minutes.