Open Workout System
Documentation for the Open Activity Library, Hardware API, and Remote Agent API. Everything is open — browse exercises, connect devices, integrate AI agents, and contribute without an account.
Activity Library
A self-governing, AI-managed exercise library. Every exercise has provenance tracking, confidence scoring, and goes through automated review.
Activity Library API
Public REST endpoints for browsing, searching, and retrieving exercises from the library.
Review Process
How the 3-tier AI review engine validates new exercises and edits before they enter the library.
Contribution Guide
How to propose new exercises, suggest edits, submit feedback, and contribute via YAML.
Hardware
Capability-based device integration. Connect smart resistance machines, cardio equipment, and sensors via a standardized descriptor and registry system.
Remote Agent API
Let external AI agents call the same tools via HTTP REST or MCP. Query training data, create exercises, log workouts, and manage programs on behalf of users via bearer token authentication.
Library Quick Start
GET /api/library/activities?limit=20 — Browse
GET /api/library/activities?q=squat — Search
GET /api/library/activities/{id} — Single exercise
GET /api/library/stats — Statistics
Hardware Quick Start
GET /api/devices/registry — All descriptors
GET /api/devices/registry?include=community — + Community
POST /api/devices/registry — Submit descriptor
Agent API Quick Start
GET /api/remote/tools — Tool manifest
GET /api/remote/context — User context
POST /api/remote/execute — Run a tool
Open source. All exercises include provenance, contributor history, and confidence scores. All device descriptors are publicly browsable. No account required to read.