%202.png)
Key Takeaways
- Sleep Score and Resilience Score now compute natively inside the platform, moving Open Wearables from pure data ingestion into Layer 2 health intelligence. Both scores are in beta.
- Every scoring algorithm is open source: auditable, forkable, and tunable for your population. Open algorithms. Zero black boxes.
- Outgoing webhooks deliver real-time events to any HTTPS endpoint, removing the need to poll for data changes. Also in beta in this release.
- The HISTORICAL_SYNC_ON_CONNECT feature flag controls historical sync behavior on new provider connections and will be removed in approximately two months.
Is Your HealthTech Product Built for Success in Digital Health?
.avif)
Open Wearables 0.4.3 shipped on April 14, 2026. The platform now computes its own health intelligence. Sleep Score and Resilience Score arrive natively, with algorithms that live in the open-source codebase. This is the first release where Open Wearables produces scored, interpreted health output from the data it already holds.
Three other features round out the release: outgoing webhooks for real-time event streaming, a grace-period feature flag for historical sync behavior, and dashboard observability improvements including an SDK logs endpoint and a seed data generator.
Health Scores (beta)
Open Wearables has always ingested wearable data from Garmin, Whoop, Polar, Oura, and the other supported providers. Starting with 0.4.3, it also scores that data.
Sleep Score evaluates sleep quality across duration, consistency, and stage composition.
Resilience Score as a single metric (HRV-CV Heart Rate Variability Coefficient of Variation) reflecting readiness and stability of recovery. This feature will be fully available in the next release.
Both scores return through the existing API alongside the underlying wearable data and can display them as-is, build custom logic on top of them, or fork the algorithm and tune it for your specific population.
The algorithms are open source. Every calculation method lives in the codebase, versioned and auditable. If a score produces a surprising result, you can trace it to the exact inputs and formulas. The platform sits between raw wearable data and the health intelligence that product teams actually need. That gap is where the scoring layer lives.
This is also where Open Wearables differs from closed-source scoring vendors. With Whoop or Oura, the recovery score arrives as a number with no method attached. With Open Wearables, every coefficient is in the repository. Fork it, tune it for your population, or contribute improvements back.
Disclaimer for 0.4.3: Sleep Score and Resilience Score are in beta. The algorithms will change as we gather feedback and improve accuracy. Do not treat beta scores as clinical-grade outputs or surface them without a disclosure that they reflect an early-stage model. Provider coverage for scoring is also limited to providers that supply the required input signals.
Outgoing Webhooks (beta)
Previous versions of Open Wearables required polling to detect new data. A sync event writes to the database; your application queries at intervals to find out.
Outgoing webhooks remove that constraint. Configure an HTTPS endpoint in your instance, register it in open wearables and events stream to your application as they occur: new activity summaries, completed sleep sessions, incoming provider push notifications. The platform sends a signed payload to your endpoint when each event happens.
Webhook payloads are signed; verify the signature before processing to confirm requests originate from your Open Wearables instance.
Historical Sync Feature Flag
A new feature flag, HISTORICAL_SYNC_ON_CONNECT, controls whether the platform initiates a historical data sync when a user connects a new provider.
The flag exists as a transition mechanism. Historical sync behavior on new connections is being standardized, and this flag lets you opt into the new behavior now or maintain the existing approach during the transition. HISTORICAL_SYNC_ON_CONNECT will be removed in approximately two months. Migrate any logic that depends on the current historical sync behavior before then. The docs include a migration guide.
We recommend setting this to False in the .env file of your Open Wearables instance so that you can start using the final version of the historical sync.
Dashboard Improvements
Two observability additions landed in the admin dashboard:
SDK Logs Endpoint: Access SDK activity logs directly from the dashboard. Useful for debugging mobile SDK integrations without dropping into server logs.
Seed Data Generator: Generate synthetic wearable data for development and testing. If you are building features that depend on health data and do not have real users connected, the seed generator produces realistic data across supported providers.
Community
As of April 15, 2026: 1,261 GitHub stars, 194 forks, 174 open issues. Seven contributors worked on this release, including two first-time contributors to the Open Wearables codebase.
What's Next
Additional score types (Strain, Recovery, Readiness), Fitbit integration, expanded scores development and the AI engine for reasoning over health scores.
Contributions are welcome. If you are interested in improving the Sleep or Resilience algorithms, the scoring code is isolated and well-documented.
Getting Started
To upgrade:
git pull origin main
docker-compose down && docker-compose up -dNew to Open Wearables? The quickstart is at docs.openwearables.io/quickstart. You can deploy it with Docker Compose or one-click on Railway. First API call in about five minutes. Self-hosted on your infrastructure, zero per-user fees, MIT licensed. If you are still evaluating your stack, the healthtech MVP stack guide covers integration options in detail.
Frequently Asked Questions
.png)
.png)
.png)
.png)
