▸Example Singular call (the "before")
import com.singular.sdk.Singular
import com.singular.sdk.SingularConfig
val config = SingularConfig("YOUR_API_KEY", "YOUR_API_SECRET")
config.withIMEICollection()
Singular.init(applicationContext, config)
Singular.event("sng_purchase", JSONObject().apply {
put("revenue", price)
put("currency", "USD")
put("user_id", userId)
})Regulated industries — telehealth, fintech, government, defence — frequently require that analytics data never leave operator-controlled infrastructure. Respectlytics's self-hosted option (AGPL-3.0 server + MIT-licensed SDKs) lets you run the entire analytics stack on your own hardware. The SDK is the same; only the API endpoint changes.
☑Remove Singular cleanly
-
1
Remove the Singular SDK from your build (
Singular-SDK/singular_sdk/singular-react-native) -
2
Remove
Singular.start(YOUR_API_KEY, YOUR_API_SECRET)andSingular.event(...)call sites -
3
Decide ATT posture — remove
NSUserTrackingUsageDescriptionif Singular was the only ATT-triggering SDK -
4
Remove the AD_ID permission from the Android merged manifest if no remaining SDK contributes it
-
5
Plan SKAdNetwork + Google Play Install Referrer as the first-party attribution replacement
⇋Singular vs Respectlytics — self-hosted deployment
| Singular | Respectlytics | |
|---|---|---|
| Self-hostable | — see tool note above (typically no) | Yes (AGPL-3.0 server) |
| Operator-managed database | — typically no | Yes (you choose Postgres deployment) |
| Air-gapped deployment possible | No | Yes |
| SDK license | — varies | MIT (permissive) |
| Server license | — typically closed | AGPL-3.0 (source-available) |
❓Frequently asked questions
What's the AGPL-3.0 obligation in practice?
If you modify the server source AND offer the modified server as a service to third parties over a network, you must publish your modifications under the same license. Internal self-hosting for your own apps does NOT trigger publication obligations. Consult your legal team for specific applicability.
What infrastructure do we need?
A Linux host (or container) running Python 3.12 + Postgres 14+ + a reverse proxy (nginx / Caddy). The default configuration scales to ~10 million events/day on a 4-vCPU / 8 GB box; larger deployments scale horizontally. See the public README for hardware sizing guidance.
Do we get updates automatically?
No — self-hosted deploys pull the latest source and run migrations on a cadence you control. We publish release notes for each version. No auto-update; you decide when to upgrade.
Are the SDKs AGPL too?
No — SDKs are MIT-licensed. Fork and modify freely with no copyleft obligation. Only the server is AGPL.