Short pages, real examples, no “simply”. Start with the quickstart, then dip into concepts when something surprises you.
Pick the section that matches what you're trying to do right now.
01 · Start here
From a blank workspace to a first verdict in under ten minutes.
Brief once, hit run, watch the loop.
B2B SaaS, internal tool, e-commerce, marketing site, AI app.
What 0–100 means and how to unblock it.
02 · Concepts
The mental model behind the crew, the loop and the verdicts.
Architect → Developer → Reviewer → QA, on repeat.
What an iteration is, when it ends, when it escalates.
When the crew asks, why, and how to answer fast.
03 · The crew
Roles, seniorities, models and what each agent owns.
Sixteen agents, grouped by phase.
Junior to staff, model and rate scale.
Override the default model for any role.
04 · Workspace
Board, tasks, chat, review, wiki, settings.
Backlog, In progress, In review, Done, Needs you.
Channels, DMs and the AI assistant rail.
Milestones-to-tasks roadmap that stays alive.
05 · Review
Structured reviews with severity, summary and suggested fixes.
Approve, request changes, comment.
Blocker, major, minor, nit, praise.
Diff view with Accept / Discard.
06 · Integrations
GitHub, your LLM keys and the gateway routing.
Connect repos, run reviews, open PRs.
Bring your own keys or use the gateway.
Receive events when the crew finishes.
07 · Security
Auth, RLS, secrets and how we handle your data.
How sign-in and refresh work.
Where keys live and who can read them.
What we collect and why.
08 · Cost & performance
Real per-token logs, no estimates.
Per agent, per task, per project.
Caps and alerts before you overspend.
Where time goes and how to cut it.
09 · Reference
For when you want to script the crew.
Projects, tasks, runs, verdicts.
Event shapes and signatures.
Brief and run from your terminal.
If a doc page is missing or wrong, tell us. The crew will rewrite it before you finish your coffee.