Journey
Built in the open, one sprint at a time.
We didn't launch Colabrix — we started shipping it. This is the arc so far, and where the work is pointed next.
Six phases, a lot of small commits, no launch event. Here's how Colabrix got here — and where the work is pointed next.
- 01
The groundwork
Rules before a line of code
A fixed stack, three ruleset files, and one way to do everything — written and committed before the first feature. No comments in the codebase, backend before frontend, agree the contract before any migration. Slow at the start so nothing had to be unpicked later.
- 02
The shape
The domain, pinned down
Organisation, project, sprint, story, task — and exactly who sees what, three levels deep, with management bypassing the gates. Argued out and written down before a single table existed, because a wrong assumption in a schema costs days.
- 03
The build
Backend, one feature at a time
Auth and sessions, orgs and invitations, billing, timed sprints, stories and tasks, comments, time logs, tags, mentions, notifications, dashboards. Each one the same way — contract, schema, service, route, docs — then the next. Every change writes to an activity log in the same transaction.
- 04
The client
A frontend from an empty folder
React from the login screen out: the workspace shell, settings, a board with drag-and-drop, backlog planning, the task panel, project and story members. Built against the running API, never ahead of it.
- 05
Live
Shipped to its own metal
One box — Postgres native, the apps and cache in Docker, a gateway in front. A pipeline behind it: work on a branch, review into staging, release on Friday. The product deploys itself now.
- 06
Where the work is
Smarter, not heavier
A per-team choice of AI engine, drafting that writes alongside you, and a read on where a sprint actually stands. The first cut is live for paid teams — and it gets sharper from here, without getting bigger.
The whole ambition: a sprint tool teams actually keep open.