Overview
Open Wearables is an open-source, self-hosted platform that unifies health data from 10+ wearable devices through one AI-ready API. Built by Momentum and MIT licensed, it ships open health scoring work — Sleep Score and Resilience Score are documented as shipped/beta in recent release materials, with broader recovery, strain, HRV, stress, and VO2 max scoring described in official marketing and roadmap materials — plus an MCP server that lets AI agents like Claude or ChatGPT reason over time-series health data without locking developers into a paid SaaS aggregator.
The platform is built for product teams who don't want to spend weeks integrating Garmin, then Whoop, then Oura one device at a time. It currently supports Apple Health, Garmin, Whoop, Oura Ring, Strava, Polar, Suunto, Samsung Health, Google Health Connect, and Ultrahuman, with iOS, Android, Flutter, and React Native SDKs covering the mobile side. Hosting it yourself means HIPAA-ready deployment and zero per-user fees — at 10,000 users, paid wearable APIs typically run $5,000–$20,000/month, while Open Wearables only costs your own infrastructure.
Key Features
- Unified Wearable API — Pulls structured time-series data from Apple Health, Whoop, Garmin, Oura, Strava, Polar, Suunto, Samsung Health, Google Health Connect, and Ultrahuman through one schema, so you write integration code once instead of per device.
- Open Health Scoring Algorithms — Provides auditable, open-source health scoring logic, with Sleep Score and Resilience Score the safest concrete examples to name today; broader recovery, strain, stress, HRV, and VO2 max scoring sit in the roadmap or marketing-positioned tier until the current release confirms each one is implemented.
- MCP Server for LLM Reasoning — Exposes wearable data to Claude, ChatGPT, and other AI assistants via the Model Context Protocol, enabling agents to answer "how was my recovery this week?" with real context.
- Health AI Engine / MCP Layer — Official materials describe trend detection, anomaly flagging, baselines, and cross-score reasoning, with availability tied to the AI reasoning layer and an evolving roadmap. Treat this as a supported and marketed capability with beta-stage caveats rather than a fully mature engine — coaching apps can still surface insights like "your HRV dropped 18% over 7 days" once the layer is wired in.
- Multi-Platform SDKs — Provides native iOS (Swift), Android (Kotlin), Flutter, and React Native SDKs for Apple HealthKit, Samsung Health, and Google Health Connect; Samsung Health access may rely on Android/Health Connect behavior and provider-specific setup constraints, with automatic background sync handled once configured.
- Self-Hosted, HIPAA-Ready Infrastructure — Runs on your own servers via Docker Compose with PostgreSQL, Redis, and Celery, giving regulated teams more control over data storage, access, audit logging, and deployment architecture. HIPAA compliance is not automatic — production use still depends on your hosting setup, security controls, policies, and any required BAA with your cloud provider or Momentum Enterprise.
How to Get Started
Open Wearables is distributed as a Docker Compose stack on GitHub, so the on-ramp depends on whether you're building solo or rolling it out to production:
- Clone and boot locally —
git clone github.com/the-momentum/open-wearables, then rundocker compose up. The team advertises "first API call in 5 minutes" once Postgres and Redis are running. - Connect provider credentials — Create OAuth apps with the wearable providers you need (Garmin, Whoop, Oura, Strava, etc.) and drop the keys into your
.envfile. Each provider has a dedicated guide in the docs. - Wire up mobile sync (optional) — For Apple Health and Samsung Health Connect, add the iOS, Android, Flutter, or React Native SDK to your app — background sync is handled automatically.
- Plug in the MCP server (optional) — Point Claude Desktop, Claude Code, or any MCP-compatible client at the bundled MCP server to give your AI assistant direct access to wearable context.
- Deploy to production — Most teams run the stack on AWS or GCP behind their own VPC; for clinical use cases, the project supports HIPAA-grade deployment with raw payload archival to S3.
Pricing & Plans
Open Wearables is fully free under the MIT License — there are no per-user fees, no usage caps tied to a SaaS bill, and no feature gating between "open source" and "paid."
Open Source (Free)
- Full access to the unified API, scoring algorithms, MCP server, and all four mobile SDKs
- Self-hosted on your own infrastructure (Docker Compose ships out of the box)
- Community support via GitHub Issues and Discord
- Unlimited users — your only cost is hosting (Postgres, Redis, app servers)
Enterprise (Custom)
- Custom managed deployment for teams that don't want to operate the stack themselves
- SLA-backed support and engineering response times
- HIPAA Business Associate Agreement (BAA) for clinical and healthcare deployments
- Pricing is contact-sales; talk to Momentum directly for a quote
The cost gap is the headline pitch: Momentum publishes that paid wearable aggregators run $5,000–$20,000/month at 10,000 users, while a self-hosted Open Wearables deployment carries only your infrastructure bill at the same scale.
Community & Ecosystem
The project sits at roughly 1,500 GitHub stars and 230 forks at the time of writing, with a steady release cadence (v0.5.0, the Product Hunt release, shipped April 29, 2026). Momentum runs the core engineering, but external contributors have already shipped meaningful integrations — the Oura Ring provider was contributed by a community member, and Fitbit, Coros, and Xiaomi support is on the public roadmap. The codebase is Python (FastAPI) on the backend and TypeScript (React + TanStack Router) on the frontend, with PostgreSQL, Redis, and Celery doing the heavy lifting; teams comfortable in those stacks can fork and extend without much ceremony. Discussion happens on GitHub Issues and the project Discord, and the themomentum_ai blog publishes architecture deep-dives roughly monthly.
Best For
- Health and fitness startups that need 5+ wearable integrations on a tight cost budget
- Teams building AI health coaches or longevity copilots that must feed structured time-series data to an LLM
- Clinical and corporate wellness products that need HIPAA-grade self-hosting and a BAA path
- Engineering teams already running FastAPI / Python infrastructure who want full control
- Open-source-first developers who refuse to lock biometric data behind a closed SaaS API
FAQ
Is Open Wearables really free?
Yes. The platform is MIT licensed and fully open-source, with no per-user fees or usage limits. You only pay for the infrastructure you self-host. Momentum offers an optional Enterprise tier with managed deployment, SLAs, and a HIPAA BAA, but pricing is custom and you only need it if you don't want to operate the stack yourself.
Which wearable devices does Open Wearables support today?
Official materials list live providers including Apple Health, Garmin, Whoop, Oura Ring, Strava, Polar, Suunto, Samsung Health, Google Health Connect, and Ultrahuman, but provider/data-type coverage varies. Fitbit, Coros, and Xiaomi are on the public roadmap. Verify the current coverage matrix in the GitHub repo before committing to a stack.
Do I need to be a Python developer to use it?
The backend is built on FastAPI (Python) and the frontend on React + TypeScript, so contributing to the core requires those skills. However, consuming the API from your own app is language-agnostic — you can call it from any HTTP client. The mobile SDKs are available in Swift, Kotlin, Flutter, and React Native.
How does the MCP server work with Claude or ChatGPT?
The bundled MCP (Model Context Protocol) server exposes your wearable data as structured context to any MCP-compatible client. Pointing Claude Desktop or Claude Code at it lets the assistant pull recovery scores, sleep stages, and HRV trends directly when answering health questions — no manual data export needed.
Is Open Wearables HIPAA compliant?
The infrastructure is HIPAA-ready, meaning self-hosting it on properly configured servers (encryption at rest and in transit, access controls, audit logging) lets you meet HIPAA technical requirements. For a Business Associate Agreement, you need either to sign one with your hosting provider (AWS, GCP) directly or use Momentum's Enterprise tier, which includes a BAA.
How much does it actually cost to run at scale?
Your only platform cost is infrastructure: a PostgreSQL database, a Redis instance, app servers, and Celery workers. Momentum publicly positions the platform as $0 per user and cites that at 10,000 users a paid aggregator API typically charges $5,000–$20,000/month — actual self-hosted spend depends on your infrastructure, retention policy, redundancy, and write volume, while hosting, database storage, LLM usage, and any managed/enterprise services remain your responsibility.
How is this different from Terra or Spike API?
Terra and Spike are paid SaaS aggregators — you pay per active user and don't see the algorithms behind their scores. Open Wearables ships the same kind of unified API but as MIT-licensed code you self-host, so you avoid per-user fees and can audit or tune the scoring math yourself. Trade-off: Terra and Spike currently support more devices out of the box.
Can I use Open Wearables in a commercial product?
Yes. The MIT License permits unrestricted commercial use, including integrating the code into proprietary apps, modifying algorithms, and building paid products on top. There's no revenue share or licensing fee, and you don't need to open-source your own application.
