Documentation
Placeholder. Full documentation is being written and lands before general
availability. The outline below shows what will live here.
What the documentation will cover
- Your first run — the intake conversation, configuring stages and tiers, the cost estimate, and the two approval gates.
- Bringing your own model key — registering an Anthropic, OpenAI or Azure AI Foundry credential, the connection check, and how key custody works (write-only: a stored key is never shown back).
- Capability tiers — what Fast, Balanced and Deep mean, and why run profiles name tiers instead of vendor models.
- The artifact set — every file a run produces, from requirements.md to the delivered archive, and how long each is kept.
- Gates, parking and resuming — what happens when a run waits for your approval, and how it continues without re-running finished work.
Curious how a finished run looks today? Browse the demo run's artifacts — it renders with no account and no model call.