@exerciseapi/mcp-server
About
Official MCP server for exerciseapi.dev — search 2,198+ exercises across 12 categories
Details
- Author
- westvegh
- Categories
- Database, Other, Search
Jump to
Setup
Install @exerciseapi/mcp-server in your MCP client (Claude Desktop, Cursor, Windsurf, and others).
Repository: https://github.com/westvegh/exerciseapi-mcp-server
Follow the installation instructions in the repository README, then restart your MCP client.
Official MCP server for exerciseapi.dev — search 2,198+ exercises across 12 categories
Official MCP server forexerciseapi.dev— give your AI coding agent direct access to 2,198+ vetted exercises across 12 categories.
Building a fitness, workout, training, rehab, or wellness app with Claude Code, Cursor, Windsurf, or Cline? Stop asking your agent to make up exercise data. Install this MCP server and your agent gets six tools that pull real, structured exercises directly from the exerciseapi.dev library.
LLMs hallucinate exercise instructions. They invent muscle groups. They mix up form cues. They'll happily tell you to do a "reverse-grip incline cable fly" with safety tips that could hurt someone.
- 2,198+ exercisesacross strength, calisthenics, yoga, pilates, mobility, physical therapy, plyometrics, stretching, and more
- Real anatomical muscle data— primary and secondary, specific muscles not just "chest" or "legs"
- Form tips and safety noteswritten by humans, not generated
- Equipment, difficulty, and category filtersso the agent can build accurate workouts
- Demonstration videosfor a growing subset of exercises (and counting)
It's a thin wrapper around theexerciseapi.dev REST API. All filtering and search runs server-side — the MCP layer just makes it callable as tools.
Claude Desktop / Claude Code— add to~/.config/claude/claude_desktop_config.json(or your project's.mcp.json):
{ "mcpServers": { "exerciseapi": { "command": "npx", "args": ["-y", "@exerciseapi/mcp-server"], "env": { "EXERCISEAPI_KEY": "exlib_your_key_here" } } } }
Cursor— add to.cursor/mcp.jsonin your project:
{ "mcpServers": { "exerciseapi": { "command": "npx", "args": ["-y", "@exerciseapi/mcp-server"], "env": { "EXERCISEAPI_KEY": "exlib_your_key_here" } } } }
Windsurf, Cline, Zed— same shape, paste into your client's MCP config.
Restart Claude Desktop / Cursor / Windsurf so it picks up the new server. You're done.
After install, ask your agent something like:
Find me five chest exercises I can do with just dumbbells, intermediate level. For each one, show me the form tips.
Your agent will callsearch_exerciseswith the right filters, get real data back, and respond with accurate, library-sourced exercises. No hallucination.
I'm building a beginner full-body workout app. Pull 8 exercises across push, pull, and legs that don't need any equipment.
What's the difference between a barbell bench press and a dumbbell bench press in terms of muscle activation?
The full tool reference is inthe MCP docs page.
Built specifically for developers using AI coding agents. The schema, the tool descriptions, the docs format — all designed so an LLM can pick this up and use it correctly the first time.
"EXERCISEAPI_KEY environment variable not set"You forgot to add the key to your MCP client config, or you didn't restart the client after adding it. Double-check the JSON in your config file and restart.
"API key is invalid"The key was rejected. Generate a new one atexerciseapi.dev/dashboardand update your config.
"Daily quota reached"You've hit the free tier's 100 req/day limit. Either wait until tomorrow (resets at midnight UTC) or upgrade atexerciseapi.dev/pricing.
The agent isn't using the toolsTwo common causes: (1) the server didn't start — check your client's MCP server logs, usually in the developer console or a log file. (2) The agent doesn't think the tools are relevant. Try asking more explicitly: "Use the exerciseapi tools to find me…"
Cold start is slow on first runnpx -y @exerciseapi/mcp-serverdownloads the package the first time you use it. To skip the download on every restart, install globally:npm install -g @exerciseapi/mcp-server, then changecommandtoexerciseapi-mcpin your config.
If you'd rather hit the REST API directly instead of using MCP — for example, you're building a server-side integration or you're not on a supported MCP client — seeexerciseapi.dev/docs. The same API key works for both.
Issues and PRs welcome. This server is intentionally a thin wrapper — if there's logic you want to change, it probably belongs in the REST API (private repo). Open an issue first if you're not sure.
Built byWest Vegh. If this saves you a weekend of building exercise data infrastructure,say hi on the dashboardor shoot me a note atwest@veghlabs.com.
Search, Query and interact with data in your Milvus Vector Database.
Connect once and your AI assistant can search and post classified listings — vehicles, real estate, jobs and CVs — on your behalf
An MCP server for vector storage and retrieval using ChromaDB.
MCP server for discovering UK regulated professional services — conveyancers, mortgage brokers, and financial advisers from SRA, FCA, and Companies House registers.
Local semantic search over documents (txt, md, pdf, docx, pptx, csv). Fully offline, multilingual, hybrid vector + keyword search via LanceDB. No API keys, no cloud.
Unified MCP server giving AI coding agents direct access to your databases: 70+ SQL databases via SqlKit (PostgreSQL, MySQL, SQL Server, SQLite, ClickHouse, Snowflake, BigQuery) and NoSQL via DocKit (Elasticsearch, MongoDB, DynamoDB). Local-first: credentials never leave your machine, read-only by default, 79 tools.
Query and export data from various databases including ElasticSearch, MySQL, PostgreSQL, Oracle, and SQLite.
Access the Discogs API for music cataloging, search, and other database operations.
Domain intelligence tools - NS reverse lookup, newly registered domain search and more
Interact with an Elasticsearch cluster via the Model Context Protocol (MCP), enabling clients to query, manage, and analyze data.
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
