LAWDE Shield diagnoses production crashes for Next.js and Node teams: it tells you what broke, why it broke, and where to fix it. Secrets are scrubbed in under 50ms, source is not stored as a durable copy, and Athena patches only clear when Aegis evidence score is at least 0.70. Start with Setup, then pick your runtime.
What is LAWDE Shield documentation for?
These pages explain how to ingest a Next.js or Node crash, run an evidence-scored diagnosis, configure .lawdeignore so sensitive paths never become AI context, and operate within tier quotas — with the same hard numbers we publish for citation: built-in scrubbing in <50ms, 0.70 trust threshold, 0 bytes durable source.
Constant
Value
Built-in privacy scrub
<50ms design target
Aegis trust gate
0.70
Durable source stored
0 bytes
Semantic clustering threshold
≥ 0.85
<50ms is our design target for the built-in privacy scrub on the hot path. Full persistence and AI clustering happen asynchronously to guarantee zero data loss.
Public Beta Launching Soon! Soft-beta users can test live error ingestion, repository-grounded AI diagnoses, and secret scrubbing on the Hobby tier.
How-to guides
Product walkthroughs for people who find the dashboard dense, and for crawlers that need citable steps. Each page matches a live screen — no mock screenshots.
LAWDE Shield diagnoses production crashes for Next.js and Node teams. It tells you what broke, why it broke, and where to fix it, scrubs secrets before persistence, and withholds a suggested patch when the Aegis evidence score is below 0.70.
Live product data always comes from your workspace database — never seed or demo crashes. Athena only runs when pipeline readiness passes (API key, real agent ingest, and GitHub Beacon) and you explicitly diagnose (or ingest queues an eligible job).
2. The 5 Core Cure Engines
Sentinel (Ingest) — Receives crash reports via HTTP or @lawdeshield/agent, authenticates x-lawde-api-key, scrubs secrets on the ingest path (built-in scrubbing in under 50ms).
Atlas (Clustering) — Groups repeating and look-alike exceptions into one issue (similarity ≥ 0.85) so you debug a failure, not a storm.
Beacon (Repository Grounding) — Uses your mapped GitHub App installation to fetch a short, read-only window of source around the crash line. Source is burn-after-read — never durably stored.
Athena (AI Root Cause & Patch) — Explains why the crash happened and suggests a fix using stack + Beacon context. Consumes daily AI-fix quota only when a diagnosis actually runs.
Aegis (Trust & Privacy Gate) — Enforces Nuclear Abort for paths matching .lawdeignore and only surfaces AI fixes at Trust Score 0.70 or above.
3. Dashboard Guide
Overview — Live clusters, topology, and first-run onboarding.
Issues — Cluster list + detail with Diagnose / Fix with Athena.
Apps — Beginner multi-language setup with your live API key filled in.
Playground — Safe test ingest (excluded from production metrics).
Profile & Settings — API keys, a ready-to-paste .env block with your live credentials, GitHub Beacon mapping, email alert preferences, and repository-synced .lawdeignore rules.
Restricted Zone — Preview Nuclear Abort path matching against your ignore rules.
4. Beginner Quickstart (test crash + live wire)
Prefer a numbered product tour? Use the guided quickstart. LAWDE does not pull crashes from your server. You push them. Complete both phases:
Follow the Central Setup Guide — copy LAWDE_API_KEY, LAWDE_WORKSPACE_ID, and optional LAWDE_APP_ID into .env.
Phase A — smoke test: use Apps → Beginner setup for pre-filled snippets (lawde-init.js, test_ingest.py, test_ingest.sh, test_ingest.js). Confirm a cluster on Overview.
Phase B — live main-app wire: keep a reporter running inside your real boot path. Pick a detailed guide:
Install the LAWDE AI GitHub App, then pick the target repository from the dropdown.
LAWDE syncs root .lawdeignore automatically when you map or re-scan the repo. Profile → Security shows Synced from repository root (.lawdeignore) when present.
Beacon is read-only and installation-scoped. If live source context cannot be fetched, Athena does not run and no AI quota is spent.
6. Tier Limits & Quotas
Plan tier
Daily ingest soft meter
Daily Athena AI fixes
Reset
Hobby
1,000 errors / day (displayed)
2 / UTC day (strict)
00:00 UTC
Pro ($29/mo)
100,000 / day (displayed)
30 / UTC day
00:00 UTC
Enterprise
Custom / Unlimited
Always-on · no hard stop
Custom
When Athena quota is exhausted, logs still land and cluster — only AI diagnosis stops until the next UTC day.
Retries do not consume extra AI diagnoses.
Daily error and AI caps are on Pricing. Burst abuse controls exist so a single noisy app cannot crowd out everyone else.
Apps: Hobby 1 · Pro up to 5.
7. Support & email alerts
Use Support & Feedback (sign-in required) or the dashboard sidebar to reach support@lawde.net. You receive an auto-confirmation.
Welcome — sent when your workspace is first created.
Crash alert — on each new unique agent cluster (throttled per cluster).
Quota reached — when Hobby hits 2/2 or Pro hits 30/30 for the UTC day.
