Why we built Regle

Between the two of us we have been on-call at companies of every size: ten-person startups where on-call meant "keep your laptop close on weekends," and ten-thousand-person platforms where on-call meant a four-person rotation with a dedicated dashboard wall and a Slack bot that ran the war room for you.

The tools at the big end of that spectrum are good. Datadog Incident Management. PagerDuty Process Automation. incident.io. They're designed for orgs that have an incident commander, a comms lead, and a procurement department to negotiate the seat license.

The tools at the small end of that spectrum are also good. UptimeRobot for a ping. Healthchecks.io for a cron heartbeat. A free PagerDuty seat to wake one person up.

And in between those — the 10-to-80-engineer band where we both spent most of our actual careers — there was nothing. So you assemble it yourself: GitHub for the deploy log, Slack for the running commentary, a Notion runbook from 2022, a feature-flag dashboard in a separate tab, and a personal mental model of which services belong to whom.

The recurring scene

Every team we've worked at, when an incident happened at this size, played out the same scene. Someone gets paged. They open seven tabs. They ask the same question — what just changed? — and they spend twenty to forty minutes answering it manually.

We kept thinking: this question is so universal, so well-defined, and so blocking to the actual work of fixing the incident, that surely somebody had built the tool. Nobody had. Or rather: people had built it inside their own companies, as an internal tool nobody open-sources because it doesn't generalize.

The wedge

So we picked it. Regle is one thing: it correlates every deploy, config change, and feature-flag flip with your incidents, and at 2 AM it tells you which change is the most likely cause. That's the whole pitch. We are deliberately not building a general observability platform. We are deliberately not building incident war-room tooling. We are not building a status page that competes with Atlassian.

We are building one feature, very well, for one buyer profile, with a five-minute setup. That's it.

The constraints we picked on purpose

• Two people, bootstrapped. No board pressure to bolt on a Fortune-500 RFP feature this quarter. We'd rather stay small and serve our band well than scale a sales motion we don't want to manage.
• Flat fee, no per-seat games. $199/month per org, unlimited users. Per-seat pricing punishes you for inviting your whole on-call rotation, which is exactly what we want you to do.
• Boring tech. Postgres, Redis, plain Go and TypeScript. We use LLMs for ranking and post-mortem drafts — not for chatbot theater.
• Founder support. When you email [email protected], one of us answers, usually within an hour during waking hours. That's a feature, not an accident. It will stop being possible at some scale and we'll deal with that when it happens.

Who Regle is for, plainly

You're a great fit if you're a platform or SRE engineer at a 10-to-80-engineer company, your team ships somewhere between 5 and 50 deploys a week, and you've had a 2 AM incident in the last month where you spent more than fifteen minutes figuring out what changed.

You're probably not a fit if you have a dedicated incident commander team, if you need ServiceNow-style change-approval workflows, or if you want a single-vendor "observability platform." We'll happily tell you so before you give us your card.

What's next

Today: alerting via webhook (PagerDuty, Opsgenie, Slack) works. Sending deploys and config changes via the same webhook or API works. The change-correlation engine works. Post-mortem drafts work.

On the roadmap: native source-control integrations (GitHub, GitLab), direct LaunchDarkly hookup, a mobile-first incident timeline view for the phone-by-the-bed case, and a small command-line tool because half of you would rather regle incidents recent than open a tab.

We're onboarding five design-partner teams this quarter — free for six months, direct line to both founders, in exchange for honest feedback. If this sounds like the tool you've been hand-assembling for years, write to us.