The conformance bar for this page: integrate from a bash script, no dependency, under 5 minutes. Everything below works from any language that can make an HTTP request.

1

Get a key

Dashboard → Add app → copy the sk_live_ server key it hands you. Or do it headlessly with the Control API: POST /v1/apps returns the key.

2

Send your first event

curl -X POST https://in.datagauge.dev/v1/events \
  -H "Authorization: Bearer $DATAGAUGE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"name":"user_signup","user":"usr_8813"}'

name is the only required field. user is any opaque string you already have: a cuid, a Firebase uid, an email, a license key. It is never format-validated.

Your app card in the dashboard flips to first event received ✓ the moment this lands.

3

Light up metrics with reserved names

`name`	Powers	Notes
`user_signup`	Signups + MAU
`payment`	Revenue	needs `props.amount` (integer cents) + `props.currency`; `sk` keys only; `id` required
anything else	Custom events, MAU	`page_view`, `report_generated`, whatever

curl -X POST https://in.datagauge.dev/v1/events \
  -H "Authorization: Bearer $DATAGAUGE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"name":"payment","user":"usr_8813","id":"order_4711",
       "props":{"amount":4900,"currency":"usd"}}'

Money is idempotent

The id makes money idempotent: send that payment twice (retries, replays, an at-least-once queue) and it counts once.

4

Add MRR with one more call

POST your subscription state whenever it changes, from your billing webhook handler, a nightly job, anywhere:

curl -X POST https://in.datagauge.dev/v1/objects \
  -H "Authorization: Bearer $DATAGAUGE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"type":"subscription","id":"sub_9","status":"active",
       "mrr":4900,"currency":"usd"}'

Upsert semantics, last-write-wins. Only status: "active" counts toward MRR. On Stripe or Polar? Skip this entirely with a webhook.

5

Read your numbers back

Wait about five minutes first

Metrics lag ingest. A read you make straight after a write returns the old value, and that is the system working, not a lost event:

Ingest batches before it stores. At the trickle volume of a first integration, a batch waits out a 30 s timeout rather than filling up.
Metric reads are cached for 300 s (5 minutes), so the first read after your write pins the pre-write answer for the rest of that window.

So budget roughly five minutes, and read again rather than concluding the event did not land. If you want to confirm the write itself immediately, your app card in the dashboard flips to first event received ✓ as soon as ingest accepts it, before any of the above.

curl https://api.datagauge.dev/v1/metrics/mrr \
  -H "Authorization: Bearer $DATAGAUGE_KEY"
# {"ok":true,"metric":"mrr","data":[{"currency":"usd","mrr":4900}]}

Also: mau, dau, signups, revenue, volume, with optional ?date_from=2026-06-19&date_to=2026-07-18.

Test it first

A DataGauge app is live by default: Add app hands you an sk_live_ key and the numbers above are real. To rehearse an integration instead, flip the dashboard to test mode (or mint an sk_test_ key) and run these same calls against it. Swap only the prefix, sk_test_ for sk_live_, and everything else is identical.

The test environment is isolated from live under its own project_id and starts empty, so that first sk_test_ event is what seeds it. Once you've sent one, test mode never shows you a blank screen. Point a real integration at sk_live_ when you're happy.

Next: the metrics reference for every metric and window, billing webhooks to get MRR without writing the objects call, or an SDK if your runtime drops background sends.

Quickstart