One number per plan.
No usage roulette.
Three plans. Flat monthly price. The error quota is the only thing that scales with you — and we'll tell you long before you hit it.
Developer
For solo projects and side hustles.
- → 1 user
- → 1 project
- → 1,000 errors / month
- → 3-day retention
- → Email notifications
- → PII scrubbing + custom rules
- → SDK version watch
- → Terminal mode
- → Community support
Basic
For small teams shipping every week.
- → Unlimited users
- → Unlimited projects
- → 50,000 errors / month
- → 1 uptime monitor
- → 14-day retention
- → Email notifications
- → AI fix prompts
- → MCP for AI assistants
- → Live log tail
- → GitHub security alerts
- → PII scrubbing + custom rules
- → SDK version watch
- → Terminal mode
- → Email support
Pro
For teams running production at scale.
- → Unlimited users
- → Unlimited projects
- → 250,000 errors / month
- → 10 uptime monitors
- → 90-day retention
- → AI fix prompts
- → MCP for AI assistants
- → Live log tail
- → GitHub security alerts
- → N+1 detection
- → Slow query analyzer
- → PII scrubbing + custom rules
- → SDK version watch
- → Integration API
- → Muhveli chat alerts
- → Audit log
- → Terminal mode
- → Email support
|
Developer
Free
|
Basic
$26/ month
|
Pro
$79/ month
|
|
|---|---|---|---|
| Users | 1 | Unlimited | Unlimited |
| Projects | 1 | Unlimited | Unlimited |
| Errors / month | 1,000 | 50,000 | 250,000 |
| Retention | 3 days | 14 days | 90 days |
| Uptime monitors | — | 1 | 10 |
| Terminal mode | Yes | Yes | Yes |
| PII scrubbing + custom rules | Yes | Yes | Yes |
| SDK version watch | Yes | Yes | Yes |
| Live log tail | — | Yes | Yes |
| AI fix prompts | — | Yes | Yes |
| MCP for AI assistants | — | Yes | Yes |
| GitHub security alerts | — | Yes | Yes |
| N+1 detection | — | — | Yes |
| Slow query analyzer | — | — | Yes |
| Integration API | — | — | Yes |
| Muhveli chat alerts | — | — | Yes |
| Audit log | — | — | Yes |
| Support | Community |
Why invite only?
We onboard in batches so the support load stays manageable and the roadmap stays honest. Drop your email below — slots open most weeks.
What counts as an error?
Every ingested event counts against your monthly quota. We don't double-bill for grouping, retries, or replays. Quota resets on the first of each calendar month.
Which platforms do you support?
Ruby on Rails — that's the first-class integration today. We use Rails ourselves and built Segfault Bin to serve Rails apps deeply, not to be a generic error tracker for everything. A JavaScript frontend integration is on the roadmap; we'll get there eventually. Other languages aren't planned.
Where is the data stored?
Everything runs on EU-region Postgres with daily encrypted backups. No US cloud, no transatlantic data-residency gymnastics.
What happens if I go over my error quota?
We keep accepting events for the rest of the day and email you before we start sampling. No surprise overage bill. Upgrade or hold tight until the next period — your call.
Can I switch plans later?
Anytime. Upgrades take effect immediately. Downgrades take effect at the end of the current billing period so your retention horizon doesn't change mid-month.
What does the slow query analyzer do?
It groups every slow statement by shape, ranks them by share of your total query time, and tells you in plain language what the shape implies — missing WHERE, unbounded COUNT, leading-wildcard LIKE, deep OFFSET, an index worth adding. It will also tell you when the answer is nothing at all and the database was simply busy, which is most of what lands in a list like this. Mute a query you've decided about and it stops being collected, not just hidden.
Can we connect our AI assistant?
Yes — Segfault Bin is an MCP server. Claude, or any MCP client, connects over OAuth and can list, read and triage issues, read logs and slow queries, and check uptime. Each connection belongs to one person and one organization, only reaches the projects that person can already open, and is switched on per member by an owner. Actions it takes are attributed to them in the activity trail. Nothing is pasted into a config file and connections are revocable from the settings page.
What happens to sensitive data in our logs?
It's filtered at ingest, before anything is written — passwords, tokens, authorization headers, session cookies, card numbers, JWTs and API-key-shaped strings, across params, headers, cookies, user objects, breadcrumbs and captured locals. That applies on every plan, free included, and you can add your own key names and regexes per project without asking us.
Do you watch our dependencies too?
If you connect GitHub, yes. Point each project at a repository and its open Dependabot alerts show up beside its errors — severity, package, CVE, the version that fixes it — with critical and high emailed to the project's recipients. It's a GitHub App, so we never hold a token: access is minted per request and you can revoke it from GitHub's side without asking us.
Ready when you are.
All three plans are invite-only for now. Tell us which one you're aiming at and we'll get you set up in the next batch.