Drengr — Understand any app. Act on any app.

# Drengr > Drengr is 0-code mobile analytics that auto-understands what your app is doing. You drop one line into the app; Drengr captures the app's real network traffic and behavior, redacts PII on-device, and auto-discovers and names the business events itself — no tracking plan, no `track()` calls. The data names itself. Drengr is built in Rust as a single binary. The Flutter SDK is available today; iOS, Android, and JavaScript SDKs are in progress. The brand is an astrolabe — a navigator's instrument for reading where you are. ## What it does - **0-code capture** — one line to install. No tracking plan, no `track()` calls, no manual instrumentation. - **Auto-discovers events** — Drengr watches the app's real network traffic and behavior and names the business events itself, so you stop guessing what to measure. - **PII redacted on-device** — customer PII is classified and sealed on the device, so it never reaches Drengr's servers readable. There is very little to comply about. - **Understanding loop** — an LLM proposes business-event mappings; a human reviews and accepts them. The LLM only names a finding once; after that, deterministic math does the counting. - **Dashboard** — funnels, retention, and failure/decline (payment declines, request failures) analysis over the events Drengr discovered. ## Why it's different - **Auto-derivation of meaning.** Many tools can see a response body. Drengr is built to auto-*understand* it — to derive and name the business event without a human writing a tracking plan first. That automatic understanding is the moat, not raw body capture. - **Verifiability over certainty.** Drengr is exact on *what happened* and honest about error bounds on what it infers. It does not promise fake-precise projections. A number you have to take on trust is worth less than one you can check. - **Math does the counting.** The LLM proposes meaning once (gated by human acceptance); deterministic aggregation handles the firehos