hevy-mcp
MCP server for the Hevy workout tracker API that allows reading and writing workouts, routines, exercise templates, and body measurements.
README
hevy-mcp
MCP (Model Context Protocol) server for the Hevy workout tracker API.
Lets Claude (or any MCP client) read and write your Hevy data: workouts, routines, exercise templates, routine folders, exercise history, and body measurements.
Requirements: Node.js >= 18, a Hevy Pro account, and an API key from https://hevy.com/settings?developer.
Usage
.mcp.json (project-level, Claude Code)
A .mcp.json file in a project's root directory makes MCP servers available to everyone who runs Claude Code inside that project — check it into git and the whole team gets the server automatically.
Step by step:
-
Create a file named
.mcp.jsonin the root of your project (same folder as.git), or run:claude mcp add hevy --scope project -- npx -y @antnsn/hevy-mcpwhich creates/updates it for you.
-
Make sure it contains:
{ "mcpServers": { "hevy": { "command": "npx", "args": ["-y", "@antnsn/hevy-mcp"], "env": { "HEVY_API_KEY": "${HEVY_API_KEY}" } } } } -
Provide the key.
${HEVY_API_KEY}is expanded from your environment at launch, so the key never lives in the file (safe to commit). Export it in your shell profile or session:export HEVY_API_KEY=your-api-key -
Start (or restart)
claudeinside the project. First use prompts you to approve the project's MCP servers; approve and thehevytools are available. Verify with/mcp.
Claude Code (global)
claude mcp add hevy -s user --env HEVY_API_KEY=your-api-key -- npx -y @antnsn/hevy-mcp
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"hevy": {
"command": "npx",
"args": ["-y", "@antnsn/hevy-mcp"],
"env": {
"HEVY_API_KEY": "your-api-key"
}
}
}
}
Global install (alternative to npx)
If npx is slow or blocked in your environment — common on corporate machines with npm security policies (e.g. allow-scripts guards) — install once globally:
npm install -g @antnsn/hevy-mcp
Then use hevy-mcp as the command in any of the configs above:
{
"mcpServers": {
"hevy": {
"command": "hevy-mcp",
"env": {
"HEVY_API_KEY": "${HEVY_API_KEY}"
}
}
}
}
Note: ${HEVY_API_KEY} interpolation only works in Claude Code's .mcp.json. Claude Desktop does not expand environment variables — put the literal key there ("HEVY_API_KEY": "your-api-key").
Tools
| Tool | Description |
|---|---|
get-workouts |
Paginated list of workouts |
get-workout |
Single workout by ID |
get-workout-count |
Total workout count |
get-workout-events |
Workout update/delete events since a date (sync) |
create-workout |
Log a completed workout |
update-workout |
Update a workout (full overwrite) |
get-routines |
Paginated list of routines |
get-routine |
Single routine by ID |
create-routine |
Create a routine |
update-routine |
Update a routine (full overwrite) |
get-exercise-templates |
Paginated exercise templates |
get-exercise-template |
Single exercise template by ID |
create-exercise-template |
Create a custom exercise |
get-routine-folders |
Paginated routine folders |
get-routine-folder |
Single folder by ID |
create-routine-folder |
Create a routine folder |
get-exercise-history |
History for an exercise template (progress tracking) |
get-body-measurements |
Paginated body measurements |
get-body-measurement |
Body measurement by date |
create-body-measurement |
Create a body measurement entry |
update-body-measurement |
Update a body measurement (full overwrite) |
get-user-info |
Authenticated user info |
Notes
- The Hevy API is v0 ("use at your own risk" per Hevy's docs) — endpoints may change.
- The API has no delete endpoints; anything created can only be deleted manually in the Hevy app.
update-workout,update-routine, andupdate-body-measurementoverwrite the full record; fetch first, modify, then update.
Contributing
From source
git clone https://github.com/antnsn/hevy-mcp.git
cd hevy-mcp
npm install
npm run build
When running from a clone, the key can also live in a .env file in the project root (gitignored) instead of the environment:
echo 'HEVY_API_KEY=your-api-key' > .env
chmod 600 .env
The repo's .mcp.json then picks the server up automatically when running claude inside the repo. Note: the .env file only works for clones — the npm-installed package looks for it next to its own install location, so use the env config shown above instead.
Development
npm run dev # run from source via tsx
npm run typecheck # type-check without emitting
Test interactively with the MCP Inspector (pass the key explicitly — the Inspector does not inherit your shell environment):
npx @modelcontextprotocol/inspector -e HEVY_API_KEY=$HEVY_API_KEY node dist/index.js
The upstream OpenAPI spec is vendored at docs-openapi.json, extracted from the Swagger UI at https://api.hevyapp.com/docs/ (the spec is embedded in swagger-ui-init.js; there is no standalone spec URL).
Issues and PRs welcome at https://github.com/antnsn/hevy-mcp.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。