Preparing your secure workspace…
Loading the latest verified data without interrupting your session.
Loading the latest verified data without interrupting your session.
2026-08-21 · Walid Abdulhameed
Onboard, send one error, see it cluster, run Diagnose, copy the agent prompt, then set scrubbers and alerts — the shortest path through LAWDE Shield.
You push a crash. LAWDE groups it, gates the AI fix at 0.70, and hands you a prompt to paste. That is the whole loop.
A live workspace, one clustered issue from a real ingest, a trust-gated diagnosis you can paste into your IDE, and a path to scrubbers and crash alerts. Deeper screens live in Docs; this is the flow.
Sign in at /signup (GitHub or Google). LAWDE creates a Hobby workspace on first login. Open Overview — it stays empty until you send an error with your API key. Copy a key from Settings → API keys (shown once).
Install @lawdeshield/agent, use Apps → Beginner setup, or POST a crash to the ingest endpoint with x-lawde-api-key. Built-in privacy scrubbing targets <50ms on the hot path so your app thread never waits on AI. Playground is a rehearsal; a real reporter in your boot path is the live wire.
Refresh Overview or open Issues. Repeats land on one card. Look-alikes join when similarity is at least 0.85. You triage one issue, not a storm of tickets.
Open the issue and click Diagnose. Connect GitHub Beacon first so Athena can borrow a short-lived, read-only source window — then drop it. That is zero source retention: we do not keep your repository. Fixes below Trust Score 0.70 stay hidden.
On a diagnosis that cleared 0.70, use Copy Agent Payload. Paste it into Cursor or any IDE. LAWDE Shield does not auto-patch your repo.
Built-in masking already strips common secrets and PII on the way in. Business+ teams add domain IDs (patient numbers, staff codes) in Settings → Custom scrubbers — a GUI, not a config file. Path protection is separate: put .lawdeignore at the repo root so sensitive files never become AI context.
Crash alerts fire when a new unique issue appears. Quota and welcome mail are automatic. Tune preferences under Profile. Daily error and AI caps live on Pricing — when AI quota is spent, logs still land; only new diagnoses wait until the next UTC day.