29MB. Full error tracking stack.

Error tracking that fits in your pocket

A single Go binary that replaces Sentry. Runs on a laptop with SQLite. Scales to Postgres when you need it. Uses your existing Sentry SDKs.

localhost:8080/issues
Issues Last 24h
247
Events
18
Issues
99.2%
Crash Free
TypeError: Cannot read prop... 142x
ConnectionRefusedError 58x
TimeoutError: deadline exceeded 47x

Pick your storage

Start with SQLite on your laptop. Move to Postgres when traffic grows. Same binary, one flag.

SQLite Mode

Zero config. Great for local dev, hobby projects, and small teams. The database is a single file you can back up with cp.

urgentry serve

Postgres Mode

For production workloads with multiple services. Concurrent writes, replication, and everything Postgres gives you.

urgentry serve --db postgres://...

What it actually costs

urgentry is free. You pay for the server. Here's what that looks like.

Laptop
Local Dev
$0 /mo

Run it alongside your app. SQLite storage.

  • Full feature set
  • SQLite storage
  • Single-user
Download
Dedicated
Scale
$20+/mo

Postgres mode. Multiple replicas if you want.

  • Full feature set
  • Postgres required
  • Horizontal ready
  • Unlimited events
Download

Meanwhile, at Sentry

Sentry's Team plan starts at $26/mo for 50K errors. urgentry has no event caps.

$5
urgentry + $5 VPS
$26+
Sentry Team plan

What's in the binary

128 Sentry API endpoints mapped. 37 already matched and working.

Error Tracking

Stack traces, breadcrumbs, user context, tags. Your Sentry SDK sends events here instead.

Performance Monitoring

Transactions, spans, distributed tracing. Find slow code without sending data off-premise.

Session Replay

See what users saw before and during a crash. DOM snapshots stored on your server.

Profiling

CPU profiles linked to transactions. Pinpoint the function, not the service.

Cron Monitors

Know when jobs don't fire. Compatible with Sentry's check-in API.

Custom Metrics

Counters, distributions, gauges, sets. Instrument business logic alongside error tracking.

Questions you're asking

Is it actually compatible with Sentry SDKs?
Yes. urgentry implements Sentry's envelope ingestion protocol. You change the DSN in your SDK config. The SDK doesn't know it's talking to a different server. We've mapped 128 API endpoints and matched 37 so far — the ones that cover errors, transactions, sessions, profiles, cron check-ins, and metrics.
What about the Sentry web UI?
urgentry ships its own web dashboard. It won't look like Sentry's UI — it's purpose-built for the data urgentry stores. If you need Sentry's full UI, Sentry self-hosted is the right call. urgentry optimizes for lightweight deployments where the full stack is overkill.
Can it handle production traffic?
We measured 183 MB/s ingest throughput on a laptop. On a $5 VPS with 1GB RAM, urgentry handles thousands of events per hour without breaking a sweat. For high-volume workloads, switch to Postgres mode and scale the database layer independently.
What happens to my data if I stop using it?
Your data. Your disk. In SQLite mode, the database is a single file — urgentry.db. Copy it, back it up, query it with any SQLite client. In Postgres mode, it's standard tables you own. There's no vendor lock-in because there's no vendor.
Why not just use Sentry self-hosted?
Sentry self-hosted requires Docker Compose, Kafka, Redis, Postgres, Clickhouse, and 2.5+ GB of RAM just to idle. That's 9+ containers. urgentry is one process, ~50 MB of memory, runs on a $5 VPS. If you need every Sentry feature, use Sentry. If you need error tracking without the infrastructure tax, use urgentry.

29 megabytes. Zero excuses.

Download it. Run it. Point your SDKs. Your errors stay on your hardware from now on.