
A minimal, self-hosted Sentry alternative for frontend error monitoring.
Capture, group, and triage runtime errors from a Next.js app — stack traces, source-mapped code
locations, distributed traces, and short session replays — without Sentry’s cost or overkill.
geniusDebug reuses the open-source @sentry/nextjs SDK pointed at its own backend; the backend
speaks the Sentry envelope protocol. No browser SDK is built. It implements only the slice of
Sentry a small team actually uses day-to-day, self-hosted on your own infra.
- Cost — no per-event / replay / seat pricing.
- Fit — own the slice you use; no vendor lock-in.
- Isolation — runs on separate infra; if geniusDebug is slow or down, the monitored app is unaffected.
- :material-rocket-launch: **[End-to-end deployment guide](/geniusDebug/deployment-guide.html)** — VPS → Coolify / AWS / DigitalOcean → project + SDK → post-deploy. Start here.
- :material-docker: **[Self-host with Docker](/geniusDebug/self-hosting-docker.html)** — one `docker compose up`, whole stack.
- :material-server: **[Deploy without Docker](/geniusDebug/deploy.html)** — Coolify (Nixpacks) or a plain VPS (pm2).
- :material-cog: **[Configuration](/geniusDebug/configuration.html)** — every environment variable, explained.
- :material-connection: **[Integrate an app](/geniusDebug/integration.html)** — point an existing Sentry app at geniusDebug.
Screenshots
Issues feed — grouped, triageable

Issue detail — symbolicated stack, highlights, trace/replay, activity

Trace waterfall

Session replay (on-error, privacy-masked)

Alerts — rules with dedupe/throttle & frequency triggers

Settings — DSN, kill switch, GitHub App, members, retention, metrics

Features
- Error grouping into deduplicated Issues with fingerprinting, short IDs, regression detection.
- Symbolication — Debug-ID → source maps in R2, original file/line/function + source context.
- GitHub — App manifest flow (personal or org), per-frame “Open in GitHub”, suspect commits, auto-resolve on
fixes SHORT-ID.
- Distributed traces — span waterfall with error markers, links back to issues.
- Session replay — on-error, privacy-masked, timeline with error markers.
- Alerts — new / regression / frequency (spike) rules, dedupe + throttle + snooze, AWS SES.
- Triage UX — filters, sort, global search (⌘K), keyboard nav (j/k/e/x/↵), merge, assign, editable highlights.
- Admin — projects, DSN keys (regenerate/revoke), members (roles), retention windows, usage stats.
- Safety — remote kill switch (disable ingest without a redeploy), back-pressure shedding, dead-letter queue.
- Scale — time-partitioned events with auto-rolled monthly partitions + retention purge.
Tech stack
NestJS · PostgreSQL (Drizzle ORM) · Redis (BullMQ) · Cloudflare R2 · AWS SES · React + Zustand +
Tailwind · @sentry/nextjs.
Status
v1 = Next.js frontend monitoring, feature-complete. Laravel/PHP is planned for v2 (SRS §12) — the
pipeline is already platform-agnostic and skips symbolication for non-JS, so v2 is client-config only.