About ExerciseBank
ExerciseBank is an MCP server for developers and workout-generator prototyping, providing access to a classified catalogue of filmed exercises and ranked substitutions.
Key features
- 01Search exercises by muscle, equipment, pattern, difficulty, and joint exclusions
- 02Retrieve exercise descriptions, coaching cues, and classifications
- 03Get ranked substitutes with scores and reasons
- 04Fetch signed video and still media URLs
Use cases
- 01Build a workout plan using exercises matched to available equipment
- 02Exclude movements that affect joints to avoid
- 03Prototype a workout generator before writing code
MCP support
Exposes ExerciseBank API tools to MCP clients, including catalogue search, exercise details, substitutions, vocabulary, and media URLs. It runs locally over stdio.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y exercisebank-mcpEB_KEYSecretYour ExerciseBank API key. Free sandbox key, no card: https://exercisebank.net/signupEB_END_USEROptional. An opaque id for the person at this machine; enables get_media_urls (signed video and still URLs). Counts as an end user on the plan.Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
4Saved tool list · updated 2w ago
This tool takes no parameters.
Details
From the project READMEexercisebank-mcp
A Model Context Protocol server for the ExerciseBank API: 1,492 studio-filmed exercises with a hand-checked classification and a ranked map of which exercises can replace which. It lets Claude, Cursor and other MCP clients search real exercises by muscle, equipment, movement pattern, difficulty and joints to avoid, so a plan is built from exercises that exist, with videos your app is licensed to show.
Use it while you build on the API (your coding agent explores the catalogue and the vocabulary as it writes your integration) or to prototype a workout generator before writing any code.
Install
Get a free sandbox key (no card) at https://exercisebank.net/signup, then:
Claude Code
claude mcp add exercisebank --env EB_KEY=eb_live_… -- npx -y exercisebank-mcp
Claude Desktop, Cursor, Windsurf and other clients (claude_desktop_config.json, .cursor/mcp.json, …)
{
"mcpServers": {
"exercisebank": {
"command": "npx",
"args": ["-y", "exercisebank-mcp"],
"env": { "EB_KEY": "eb_live_…" }
}
}
}
It runs on your machine over stdio; the key stays in your client's configuration. Node 18 or later.
Tools
| Tool | What it does |
|---|---|
get_vocabulary | The exact filter values: muscles, equipment, movement patterns, joints, impact; the catalogue version; the ids a sandbox key gets video for |
search_exercises | Search and filter: q, muscle, equipment (the gear the person has), pattern, minDifficulty, maxDifficulty, excludeJoints, impact, unilateral, limit, offset |
get_exercise | The full record: description, coaching cues, French, classification |
get_substitutes | Ranked replacements with a score and a reason each, filtered by gear and joints to avoid |
get_media_urls | Signed video and still URLs. Only when EB_END_USER is set (below) |
Video
ExerciseBank signs media only for a named person, and that person counts as an end user on the plan.
To get get_media_urls, add an opaque id for whoever is at this machine:
"env": { "EB_KEY": "eb_live_…", "EB_END_USER": "dev-laptop-anna" }
URLs expire after 15 minutes. A sandbox key gets media for the 50-exercise sample listed by
get_vocabulary; a paid plan gets all 1,492.
In your product
This server is for the developer's desk. A product calls the API from its own backend, with its own
users' ids: use the exercisebank client and see the
guide, an AI workout generator with Claude.
- API reference: https://exercisebank.net/docs · for language models: https://exercisebank.net/llms.txt
- Pricing, by monthly end users: https://exercisebank.net/pricing
MIT licensed. The library itself is licensed through a plan: https://exercisebank.net/license
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.