Getting started
From download to a first agent run, in about four minutes.
This documents the macOS private beta. Until a verified release manifest is published, these steps describe how Codoro works rather than something you can install today.
1 — Connect a model
On first launch Codoro asks for a model connection. Provider keys go to the macOS keychain. Hosted models use your Codoro sign-in and never expose Codoro's provider credentials to the desktop.
2 — Set your caps
A per-run and a per-day cap are required before the first agent run. They can be changed at any time, and on a Team plan the owner sets a ceiling that a member cannot raise past.
3 — Open a repository
Codoro searches locally on demand. Only the ranked excerpts a request needs are sent to the selected model, and the Plan board reads git history directly.
4 — Run your first task
Describe the change in the Agent panel. A governed planning call produces staged file diffs and any allowlisted build or test checks. Approving writes that exact plan; optimistic undo restores the reviewed source files unless newer work has changed them.
5 — Design a screen
The Design tab shows manifest-declared frame previews, each one bound to a repository path. Select a frame and the properties panel edits the tokens the app ships — the hairline gap, the zero radius, the 1240 maximum width — read from app/globals.css rather than copied into a second document. Apply to code quotes an estimate, stages the change as a diff, and commits nothing until you approve it.
6 — Talk to it
Voice shows microphone permission and level state and can narrate a reviewed diff with local system speech. Speech-to-text controls appear only on targets where recognition is actually available.
Shortcuts
| Keys | Action |
|---|---|
| ⌘K M | Switch model, keeping context |
| ⌘⏎ | Run the current step |
| ⌘⇧Z | Undo the entire run |
| ⌘⇧P | Plan board |
| ⌘⇧C | Team chat |