Analysis updated 2026-05-18
Sync your own Fitbit or Pixel health data locally instead of using a paid cloud service.
See transparent, explainable readiness, sleep, and training load scores instead of a black box number.
Chat with an AI coach that answers questions about your synced health data using real charts.
Get a daily briefing and weekly retrospective summarizing changes in your health metrics.
| deekshith-dade/fettle | 0xustaz/streamgate | a-bissell/unleash-lite | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | Python | Python | Python |
| Setup difficulty | hard | hard | hard |
| Complexity | 4/5 | 4/5 | 4/5 |
| Audience | developer | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires a Google Health API OAuth setup, a Fitbit or Pixel device, and a logged-in opencode CLI for free AI model access.
Fettle is a personal health dashboard that lets you own and explore your Fitbit or Pixel watch health data instead of relying on a paid subscription service. It connects to Google's Health API, pulls your data down into a local database file on your own machine, and then calculates its own transparent versions of readiness, sleep, and training load scores, showing exactly how each number is calculated rather than hiding the formula. The dashboard covers an overview page with a daily readiness score and your goals, plus deeper sections for sleep analysis, workouts, comparisons against general population benchmarks, and a drill down view for every type of metric it syncs, from daily step counts to minute by minute heart rate during workouts. It also runs an insights engine that looks for trends, unusual spikes, training load imbalances, sleep debt building up over two weeks, and links between different metrics, always describing these as associations rather than claiming one thing causes another. A notable feature is a built in AI coach you can chat with about your own data. It can pull up charts, compare stats, and explain your recent workouts or sleep, and it remembers ongoing details you mention, like an injury or an upcoming event, bringing them into later conversations. After each sync it also writes a short daily summary of what changed, and on Sundays a weekly review. Notably, the coach does not do any of its own math. It calls a fixed set of tools that read from the same deterministic engine used elsewhere in the app, so the numbers it discusses are always the same ones computed by the dashboard itself. The project runs on a Next.js frontend and a FastAPI backend, using a command line AI tool the user is expected to already be logged into for free language model access, so no separate AI subscription cost is required. It is built for one person's personal use, syncing only your own account, and is not intended to be run as a shared or public service.
A self hosted dashboard that syncs your own Fitbit or Pixel health data locally and adds a transparent readiness engine plus a free AI health coach.
Mainly Python. The stack also includes Python, FastAPI, Next.js.
The explanation does not mention a license, so terms of reuse are unclear.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.