Wingtics separates the analytics dashboard from the tool that does the measuring. You give each user a connector for the provider they already use and components that render the data. Moving from Plausible to GA4, Vercel, Umami or PostHog means swapping the connector: the charts stay the same.
The problem it solves
Every analytics tool has its own API, its own metric names and its own limits. Build a stats panel on top of one and you are locked in: switching providers means rewriting the panel. Wingtics puts a shared query model in between, so the panel neither knows nor cares where the data comes from.
What I built
- A canonical query model (
@wingtics/core): metrics, dimensions and ranges with the same names for every provider, with caching built in. - Five connectors: Plausible, Vercel Web Analytics, Google Analytics 4, Umami and PostHog, plus a mock connector for demos and tests.
- React widgets and a dashboard (
@wingtics/react), also available as a shadcn registry. - Next.js handlers (
@wingtics/next) that proxy requests, so API keys never reach the browser.
Key decisions
Each connector declares its capabilities: which metrics and dimensions it can answer. A widget that needs bounce rate, wired to a provider that does not offer it, shows "Not supported" instead of a zero that looks like real data. An honest gap beats an invented number.
Adding a provider or a widget does not touch the query model: a provider implements defineConnector and a widget implements defineWidget.
Results
- Nine packages published on npm, MIT licensed.
- Live demo, docs and a component catalog at wingtics.com.
Project link: wingtics.com
