You want to know what improved over the last six weeks—not just view another workout list.
Connect Hevy to the AI assistant you already use
Talk to your Hevy data from your AI assistant.
Hevy MCP is the connector that lets Claude, Cursor, Codex, and other AI assistants use your workouts, routines, exercise history, and measurements. Analyze progress, find what you need, and log changes in plain language.
Independent, open-source connector for the Hevy API.
25 focused tools3 connection methods18-second demoMIT licensed open source
The problem
Your Hevy data is useful. Getting answers from it should be easier.
Your training history contains the evidence, but answering a multi-part question still means opening old sessions, comparing numbers, and searching through routines manually.
You remember a routine or exercise but cannot remember where it lives in Hevy.
You want to create or update a workout without inventing missing training details.
Instead of opening and comparing individual workouts, ask multi-week questions and receive an answer grounded in specific sessions.
What you can do
Turn your training log into a useful conversation.
01
Understand your progress
Summarize 1–12 weeks of workouts and measurements to see progression, consistency gaps, and the sessions behind the trend.
Load ↑Effort ↓Evidence: 5 sessions
02
Find routines and exercises faster
Search your Hevy history, routines, folders, and exercise catalog in plain language. Get compact results instead of oversized responses.
push day3 results
Push Day A6 exercises · Chest, shoulders, triceps
Push / Pull8 exercises · Upper body
Travel Push4 exercises · Dumbbells
03
Plan and log training
Create or update workouts, routines, custom exercises, and measurements from a conversation. Compatible clients can request confirmation before writing changes.
ExerciseSetsTop set
Incline bench press355 kg × 10
Chest press machine365 kg × 9
Cable lateral raise412.5 kg × 14
Mutation requires confirmationReady
25 focused toolsExplore the complete MCP tool manifestView tools
Analysis
get-training-summary
Workouts
get-workoutsget-workoutget-workout-countget-workout-eventscreate-workoutupdate-workout
Routines
search-routinesget-routinesget-routinecreate-routineupdate-routineget-routine-foldersget-routine-foldercreate-routine-folder
Exercises
get-exercise-templatesget-exercise-templatesearch-exercise-templatescreate-exercise-templateget-exercise-history
Measurements
get-body-measurementsget-body-measurementcreate-body-measurementupdate-body-measurement
Account
get-user-info
Recorded demo · 18 seconds
Watch the evidence-gathering flow at full size.
- Retrieves real Hevy workout history
- Grounds answers in specific sessions
Full-resolution recording
Watch the recorded session Open the 18-second session in your browser's video player. The page keeps the readable proof in the interactive example above.
Setup
Connect Hevy.
Ask your first question.
Start with browser sign-in for the shortest path to all 25 tools. Prefer a fixed API key or local control? Direct HTTPS and local mode are available.
Choose how your assistant connects. You can switch methods without losing the instructions below.
Browser sign-in is recommended for most people. Use direct HTTPS when your client accepts a fixed API key, or local mode when you want the process on your machine.
01
Get and verify your Hevy API key
Create a Hevy API key and keep it handy. Browser sign-in asks for it once and stores it encrypted in your OAuth grant; direct HTTPS sends it as a bearer credential; local mode passes it through your client's environment.
Open Hevy API settings Hevy PRO is required for API access.
Check the key before configuring your clientThis makes a read-only account request directly to Hevy.
For this check only, your browser sends the key directly to Hevy. This site never receives or stores it.
02
Choose your AI assistant
Add the hosted endpoint and your assistant opens a secure browser flow. There are no environment variables, JSON files, or local server processes to manage.
https://hevy.chrisdoc.dev/mcpNextAdd this URL as a custom connector. Continue in the browser, paste your Hevy API key once, and approve access.
https://hevy.chrisdoc.dev/mcpNextAdd this as a remote MCP server. Compatible OAuth 2.1 clients discover the authorization flow and open it in your browser.
02
Choose your AI assistant
Nothing to install or keep running. Connect your client to the hosted endpoint and provide your Hevy key as a bearer credential.
export HEVY_API_KEY=your-hevy-api-key
codex mcp add hevy \
--url https://hevy.chrisdoc.dev/mcp \
--bearer-token-env-var HEVY_API_KEYBefore you save: replace your-hevy-api-key and never commit or screenshot the real value.
After savingRestart Codex or begin a new session, then run codex mcp list to verify the server is configured.
{
"mcpServers": {
"hevy": {
"url": "https://hevy.chrisdoc.dev/mcp",
"headers": {
"Authorization": "Bearer your-hevy-api-key"
}
}
}
}Before you save: replace your-hevy-api-key and never commit or screenshot the real value.
After savingRestart or reconnect the client. It must support Streamable HTTP and a fixed Authorization header.
02
Choose your AI assistant
Run the same server locally when your client cannot send a fixed authorization header or you prefer to control the process on your machine.
codex mcp add hevy \
--env HEVY_API_KEY=your-hevy-api-key \
-- npx -y hevy-mcpBefore you save: replace your-hevy-api-key and never commit or screenshot the real value.
After savingRestart Codex or begin a new session, then run codex mcp list. Local npx usage requires Node.js 20 or newer.
{
"mcpServers": {
"hevy": {
"command": "npx",
"args": ["-y", "hevy-mcp"],
"env": {
"HEVY_API_KEY": "your-hevy-api-key"
}
}
}
}Before you save: replace your-hevy-api-key and never commit or screenshot the real value.
After savingSave the client configuration, then restart or reconnect the client. Local npx usage requires Node.js 20 or newer.
npx -y hevy-mcpAfter savingConfigure the client to launch this command with HEVY_API_KEY in the child-process environment, then reconnect.
03
Ask your first question
After reconnecting your client, start with this read-only question:
“Which Hevy account is connected?”
If your assistant returns your Hevy profile, the connection is working. Then try: “Give me a training summary for the last four weeks.”
Questions, answered
Know what you are connecting before you start.
The short version: browser sign-in is the fastest path, Hevy PRO is required for API access, and you can run the open-source server locally when you prefer.
What is Hevy MCP?
Hevy MCP is an open-source connector that lets Claude, Cursor, Codex, and other MCP clients read, analyze, create, and update data in your Hevy account.
Do I need Hevy PRO?
Yes. A Hevy API key is required, and API access currently requires a Hevy PRO subscription.
Do I need to install anything?
No for browser sign-in. The hosted endpoint requires no Node.js, Bun, Docker, environment variables, or local server process. Local mode uses npx, Bun, or Docker; local npx usage requires Node.js 20 or newer.
Which assistants work with Hevy MCP?
The setup includes paths for Claude.ai, Codex, Claude Desktop, Cursor, and other compatible MCP clients. The available connection method depends on whether your client supports browser OAuth, Streamable HTTP, or local stdio.
Is my Hevy API key stored?
The optional browser key check sends your key directly to Hevy and does not send it to this landing site. Browser sign-in validates the key and stores it encrypted inside the OAuth grant. Direct HTTPS validates the key per request without storing it. Local mode passes the key through your local process environment.
Can Hevy MCP change my workouts?
Yes. It can create and update workouts, routines, custom exercises, and body measurements. Compatible clients can request confirmation before mutation tools run, so review the proposed change before approving it.
Is this an official Hevy product?
No. Hevy MCP is an independent open-source project and is not affiliated with or endorsed by Hevy.
Start with one useful question
Put your Hevy data to work.
Connect in the browser for the shortest path—no local install required. Prefer local control? Run the same open-source server on your machine. Hevy PRO API access is required. Inspect the source on GitHub and choose the connection method that fits your workflow.