In active development · MIT open source
Free, self-hosted error tracking for your Sentry SDKs.
Free, MIT-licensed server and web app. Keep your Sentry SDKs, change one line: the DSN.
docker run -d -p 8080:8080 -v openfault:/data ghcr.io/openfault/openfault 01 · Drop-in
Keep your SDKs. Change one line.
OpenFault speaks the Sentry envelope protocol. The official Sentry SDKs post to it as they are. No fork, no plugin, no wrapper.
- Rails
- Node
- Next.js
Other Sentry SDKs post the same envelope.
Minified stack traces? sentry-cli sourcemaps upload works against it, same flags, same debug IDs.
iOS crashes? Upload dSYMs from an Xcode build phase. Swift frames come back demangled, with file and line. sentry-cli debug-files upload works too, in chunks, with no size cap.
Not doing Tracing, session replay, profiling. On purpose. OpenFault tracks errors. That is the whole job.
sentry.client.config.ts
Sentry.init({
- dsn: "https://[email protected]/56",
+ dsn: "https://[email protected]/1",
tracesSampleRate: 0,
}); 02 · Grouping
Grouping that doesn’t split bugs.
Issues are keyed on the shape of the stack: module, file, function. Never line numbers. Add a log line above the crash and the next event still lands in the same issue.
A second key on exception type and message catches refactors that move the stack. Want your own rule? Send a fingerprint. {{ default }} keeps our key in the mix.
Search like is:unresolved release:1.2 assignee:me, save it as a view, and act on many issues at once.
order.rb · Order#total
41 def total42 logger.info("total")43 items.sum(&:price)raises44 end- #4812orders | order.rb | total2 evtevt a1f3order.rb:42euevt 9c07order.rb:43eu
Event 9c07 from order.rb line 43 grouped into issue #4812, 2 events.
03 · MCP
Built for agents. Gated by people.
MCP is built in. Every tool calls the same handler the UI uses, so an agent sees what you see and nothing more.
An agent can list, search and triage. It can’t read the stack or start a fix until a person clicks Approve. Captured data is untrusted: every message, breadcrumb and tag comes back fenced, with a note telling the model not to follow it.
- list_new_issues
- search_issues
- get_issue
- resolve_issue
- mute_issue
- link_fix
Admin tokens add admin tools: create projects and teams, invite members. Any issue reads as Markdown, ready to paste into a prompt.
04 · One container
One container. One volume. That’s the stack.
A Rust binary and SQLite. No Redis, no Kafka, no Postgres, no ClickHouse. It runs next to your app on the small VPS you already pay for. IPv4 and IPv6 out of the box.
docker run -d -p 8080:8080 -v openfault:/data ghcr.io/openfault/openfault
docker stats
ls /data
- Image
- ~10 MB download
- Peak RAM, 10k-event burst
- 28 MB measured
- Default retention
- 90 days
- Backup
- /data Copy the volume, or back up daily to S3 or R2.
05 · Live
Live, on the wall.
Counters, rows and charts update live as errors land. No polling, no refresh.
Sign in once, open /tv and press F for fullscreen. It shows big numbers, trends and the latest issues.
06 · iOS On the App Store
In your pocket.
Pair it in a minute: scan the QR code in Settings → Devices.
Push with Resolve and Mute right on the notification.
Home and lock screen widgets for unresolved counts and the latest issue.
07 · Teams
Teams and alerts.
It’s for the whole team, not just whoever set it up.
- Invites and roles
- Owner, admin, member. Group people and projects into teams. Invite by email or send a link. Invite only. There is no public signup page.
- Alerts
- On new issues, regressions and spikes. Per person, per project.
- Channels
- Email, Slack, Discord, Teams, Mattermost, Telegram, or any webhook signed with HMAC.
08 · Your data
How it handles your data.
Everything in an event is untrusted. OpenFault scrubs it before it is stored and fences it on the way out.
- Scrubbed on ingest
- Passwords, tokens, cookies and keys become
[Filtered]before anything is written. Values are checked for bearer tokens, AWS and GitHub tokens, Stripe keys and card numbers. Raw payloads are never logged. - IP addresses
- Dropped unless the project opts in.
- On your server
- One SQLite file in
/data. Passwords use Argon2. Sessions, API tokens and invite codes are stored only as hashes. - Outbound
- Only for webhooks, email and the push relay. Webhooks are
https://only and never reach private addresses. Source maps are never fetched from URLs. - Agents
- Stack frames, breadcrumbs and request data stay hidden from tokens until a person approves the issue.
09 · Compare
How it stacks up.
Measured, not promised: 10 MB RSS idle, peak 28 MB in a 10k-event burst across 3 projects (3,000 accepted, 7,000 rate-limited; tests/load.rs) (Apple M5, macOS 27, 2026-11-06).
Full Sentry comparison → · All alternatives → · Sentry self-hosted requirements →
| OpenFault | Sentry self-hosted | GlitchTip | Bugsink | Telebugs | |
|---|---|---|---|---|---|
| License | MIT | FSL (converts to Apache 2.0 after 2 years) | MIT | PolyForm Shield 1.0.0 (source-available, not OSI-approved) | Proprietary |
| Self-host price | Free, MIT, no paid tier | Free | Free | Free | $299 first year, $99/yr renewal |
| Hosted plan | None | Developer free (1 user, 5k errors); Team $26/mo annual, $29 monthly | Free up to 1k events/mo; $15/mo small plan | Free 15K-event evaluation plan; paid from €16/mo (checked 2026-11-09) | Managed $99/mo after a 14-day trial (checked 2026-11-09) |
| Containers to run | 1 | Many (Docker Compose stack) | App + Postgres | 1 | 1 |
| Database | SQLite | Postgres, ClickHouse, Kafka, Redis | Postgres | SQLite supported | SQLite |
| Minimum RAM | Peak 28 MB in a 10k-event burst (measured) | 16 GB + 16 GB swap | 256 MB minimum, 512 MB recommended | — | 1 GB |
| Runtime | Rust + SQLite | Python, Kafka, ClickHouse, Postgres, Redis | Python (Django) | Python | Ruby on Rails |
| Sentry SDK compatible | Yes | Yes | Yes | Yes | Yes |
| Source maps | Yes, via sentry-cli | Yes | Yes, via file upload | Yes, via sentry-cli | Yes |
| iOS crash symbolication | Yes, dSYM upload + Swift demangling | Yes | — | — | — |
| Live updates and TV mode | Yes, SSE | — | — | — | — |
| MCP server | Built in, approval-gated | Separate server | Built in, enabled by env var (per glitchtip.com docs, checked 2026-11-21) | — | Built in (per telebugs.com, checked 2026-11-20) |
| iOS app | Widgets, push, QR pairing (on the App Store) | — | — | — | — |
— means we have not checked.
Verified 2026-11-21. Sources: Sentry self-hosted, Sentry pricing, GlitchTip, Bugsink, Telebugs.
10 · Get started
Get started in minutes.
One command, no config. Everything lives in /data. Open http://localhost:8080 and the setup wizard creates your admin account and first project.
-
Run the container.
docker run -d -p 8080:8080 -v openfault:/data ghcr.io/openfault/openfault Change the DSN. Copy the project’s DSN into your Sentry SDK config and deploy.
-
Trigger a test event. No app handy? Send one by hand with the public key from the DSN.
curl http://localhost:8080/api/1/envelope/ -H 'X-Sentry-Auth: Sentry sentry_key=<public-key>' --data-binary $'{}\n{"type":"event"}\n{"message":"hello from OpenFault","level":"error"}\n'
11 · FAQ
Questions.
Is it really drop-in?
For errors, yes. OpenFault accepts the same envelope endpoint the Sentry SDKs already post to. Point the DSN at your server and events land. Items we don’t handle, like transactions, are dropped, not rejected, so your app never notices.
Is there a mobile app for error alerts?
Yes. The OpenFault iOS app is free on the App Store, native Swift, closed source. The server, relay and web app are open source (MIT). The iOS app is closed source for now. Pair it by scanning the QR code in Settings → Devices. Pushes arrive with Resolve and Mute right on the notification, and home and lock screen widgets show unresolved counts and the latest issue. Push is off until you opt in, and then works through the relay configured in Settings. A hosted push relay is available on request, free during the beta. The app needs a running OpenFault server.
Does it work with AI coding agents?
Yes. OpenFault is an MCP server. Claude Code, Cursor or any client that speaks Streamable HTTP can list, search and triage issues, and can set up a fresh instance end to end. A person clicks Approve for fix before an agent gets an issue’s stack frames, breadcrumbs and request data; nothing starts on its own. Event data is wrapped as untrusted so the agent does not follow instructions hidden in it. Give agents a read token unless you want them to resolve or mute issues. Once an issue is approved, an agent can attach the pull request that fixes it.
What about performance tracing?
Not planned. No tracing, no session replay, no profiling. If you need those, keep a tool for them and send errors here. Doing one job is how this stays one small container.
Can I migrate from Sentry?
Change the DSN in each app and deploy. New errors flow into OpenFault from then on. We don’t import history, so keep Sentry around read-only until you stop needing the old issues.
How is data retained?
Events are kept for 90 days by default, set per project. Within an issue we keep the first 25 events of each hour, then 1 in 20, and the counters always count. A noisy bug can’t fill your disk. Passwords, tokens and cookies are changed to [Filtered] before anything is stored. Everything lives in /data.
Is it production ready?
Not yet. OpenFault is in active development and nothing here is ready for production. Follow the repo on GitHub to see when the first release ships.
Self-host it in a minute.
Free, MIT-licensed server and web app. Keep your Sentry SDKs.