hevy-mcp
Provides tools for reading a user's Hevy training data — workouts, workout events (incremental sync of updates and deletes), routines, routine folders, exercise templates, and exercise history — and for managing routines by creating and updating them through the Hevy API.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
hevy-mcp
A small, self-hosted MCP server on Cloudflare Workers that lets Claude read your Hevy training data and manage routines through a custom connector.
Reads: workouts, workout events, routines, routine folders, exercise templates, exercise history
Writes: routines only (create and update). Workouts are read-only in v1.
No database, no Durable Objects, no OAuth. One Worker, two secrets.
How it works
flowchart LR
C[Claude custom connector] -->|HTTPS POST /mcp/<token>| W[Cloudflare Worker]
W -->|token mismatch| X[404]
W -->|token ok| H[createMcpHandler<br/>agents/mcp/server]
H --> S[McpServer<br/>12 tools]
S -->|api-key header| A[Hevy API<br/>api.hevyapp.com/v1]The Worker is stateless: every request builds a fresh McpServer, checks the path token in constant time, and proxies tool calls to Hevy with your API key. Anything that is not /mcp/<token> gets a 404, so the URL itself is the only credential a client needs.
Related MCP server: hevy-mcp
Tools
Tool | Hevy endpoint | Notes |
|
| Max 10 per page |
|
| Full sets and exercises |
|
| |
|
| Incremental sync of updates and deletes |
|
| Max 10 per page |
|
| |
|
| Max 10 per page |
|
| Write |
|
| Write, replaces the whole routine |
|
| Max 100 per page |
|
| |
|
| Optional |
Routine write bodies follow the Hevy spec exactly. Set fields are type (warmup, normal, failure, dropset), weight_kg, reps, distance_meters, duration_seconds, custom_metric and rep_range. There is no rpe field on routine sets.
Setup
Requirements: Node 22+, a Cloudflare account, and a Hevy Pro subscription (the API key lives at hevy.com > Settings > Developer).
npm install
npx wrangler login
# Secrets never go in code. Set them on the Worker:
npx wrangler secret put HEVY_API_KEY
npm run token # prints a random path token
npx wrangler secret put MCP_PATH_TOKEN
npm run deployAfter deploying, note the workers.dev URL that wrangler prints. Verify the gate:
curl -i https://hevy-mcp.<your-subdomain>.workers.dev/ # expect 404
curl -i -X POST https://hevy-mcp.<your-subdomain>.workers.dev/mcp/<token> \
-H 'content-type: application/json' -H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl","version":"0"}}}'The second call should return a serverInfo block naming hevy-mcp.
Connect it to Claude
In Claude, open Settings > Connectors > Add custom connector.
Name it Hevy and paste
https://hevy-mcp.<your-subdomain>.workers.dev/mcp/<token>as the URL.Leave authentication empty. The token in the path is the credential.
Start a chat, enable the connector, and ask for your last three workouts. Compare against the app.
Test a write by asking Claude to create a throwaway routine, confirm it appears in Hevy, then delete it in the app.
Treat the full URL like a password. If it leaks, run npm run token, set the new value with wrangler secret put MCP_PATH_TOKEN, redeploy, and update the connector.
Local development
cp .dev.vars.example .dev.vars # fill in real values; the file is gitignored
npm run dev # http://127.0.0.1:8787/mcp/<token>
npm run check # typecheck + wrangler dry-run buildProject layout
src/index.ts Worker entry, token gate, McpServer factory and all tool definitions
src/hevy.ts fetch wrapper for api.hevyapp.com/v1 with typed errors
wrangler.toml Worker config (nodejs_compat, no bindings)Design notes
Built on
createMcpHandlerfrom theagentspackage rather thanMcpAgent. The latter is now marked deprecated and feature-frozen upstream, and it needed a Durable Object that a read-mostly proxy never used.The first deployment of this Worker did use a Durable Object class.
wrangler.tomlcarries adeleted_classesmigration so redeploying cleans it up. On a brand new Worker the two migrations cancel out.Write tools stay limited to routines until reads have been trusted for a while. Adding workout or body-measurement writes is a matter of registering more tools against endpoints that already exist in the Hevy spec.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Create Hevy routines and analyze your training from chat. Unofficial; BYO Hevy PRO API key.
Connect Claude to your Intervals.icu watch data for fitness, workout review, and plan writing.
Pace is a remote MCP server that exposes wearable and fitness data to Claude via the Model Context Protocol. It connects to Garmin, Oura, Whoop, Polar, Fitbit and 20+ devices and provides 15 tools for querying sleep, activity, recovery, and training data. Hosted on Google Cloud Run, OAuth 2.1 authentication, Streamable HTTP transport. Instructions: First you need to create an account at: https://pacetraining.co and connect your wearables. After that you can connect the remote Server via Custom Connector in Claude and OAuth 2.1 Flow startet.
Your Garmin, Oura, WHOOP, Hevy and 35 more fitness apps, queryable from your own AI.
Related MCP Servers
- FlicenseNot gradedqualityBmaintenanceEnables AI assistants to interact with the Hevy fitness tracking API for logging workouts, managing routines, and tracking fitness progress.34-
- AlicenseNot gradedqualityDmaintenanceExposes the Hevy workout API to Claude, enabling users to manage workouts, routines, exercise templates, body measurements, and user info via natural language.MIT
- FlicenseNot gradedqualityDmaintenanceEnables AI assistants to interact with the Hevy fitness tracking API, allowing users to log workouts, manage routines, browse exercises, and track fitness progress through natural language.-
- FlicenseNot gradedqualityDmaintenanceEnables LLMs to read and write Hevy workout tracking data, including workouts, routines, exercise templates, per-exercise history, and body measurements through all 15 Hevy public API endpoints.-