Free, privacy-first telemetry for open-source projects.

Most analytics platforms are overkill (and expensive) for the kind of usage data open-source maintainers actually need: which versions are people running, what features get used, where do crashes happen? statsfactory gives you that insight without costing a cent -- it deploys as a single Cloudflare Worker on the free tier with D1 for storage. No servers to manage, no monthly bills, no vendor lock-in.

• Zero cost -- runs entirely within Cloudflare's free tier (Workers + D1). Comfortable for thousands of events per day across multiple projects.
• Privacy-first -- no cookies, no fingerprinting, no PII stored. Events carry only the dimensions you explicitly send.
• Multi-dimension analytics -- cross-tabulate any combination of dimensions in one view (e.g. plugin.name x plugin.version x status).
• Drop-in SDKs -- lightweight clients for TypeScript/JavaScript and Go, or use plain HTTP POST.
• Self-hosted, your data -- deploy to your own Cloudflare account in minutes. You own the data, full stop.

Prerequisites: bun, a Cloudflare account with a domain. See docs/deploy.md for full prerequisites and details.

git clone https://github.com/jmylchreest/statsfactory.git
cd statsfactory
./deploy.sh install

The script handles everything interactively: D1 database, domain, Cloudflare Access, build, and deploy. Then create an app and key in the dashboard and start sending events:

import { StatsFactory } from "@statsfactory/sdk";

const sf = new StatsFactory({
  serverUrl: "https://stats.example.com",
  appKey: "sf_live_xxxx",
  clientName: "myapp",
  clientVersion: "1.0.0",
});

sf.track("plugin_used", {
  "plugin.name": "kitty",
  "plugin.version": "0.1.27",
  "plugin.status": "ok",
});

No SDK? The full API is described by an OpenAPI 3.1 spec at GET /v1/doc on your deployed instance, or use curl directly:

curl -X POST https://stats.example.com/v1/events \
  -H "Authorization: Bearer sf_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"events": [{"event": "app_started"}]}'

Other deploy commands:

./deploy.sh upgrade              # Apply migrations, rebuild, redeploy
./deploy.sh reconfigure-access   # Change who can access the dashboard
./deploy.sh destroy              # Tear everything down

Prerequisites: bun, just

just init   # install deps, apply schema, seed test data
just run    # start wrangler dev at http://localhost:8787

The seed output prints a test app key (sf_live_...). Auth is bypassed automatically in local dev.

apps/web/          Cloudflare Worker (Hono API + Astro dashboard)
packages/sdk-go/   Go SDK
packages/sdk-ts/   TypeScript SDK (browser, Node, Bun, Deno, Workers)
docs/              Deployment and operations docs

How is this different from Aptabase / PostHog / Plausible? Purpose-built for the open-source use case: deploy once for free, send events with typed dimension maps, and cross-tabulate any combination of dimensions.

What does "free" actually mean? Cloudflare Workers free tier: 100K requests/day. D1 free tier: 5GB storage, 5M reads/day, 100K writes/day. Enough for thousands of telemetry events per day. The Workers paid plan ($5/month) removes all daily caps.

MIT