Skip to content

Cookbook

Safe existing-project adoption

ai-workspace doctor --verbose
ai-workspace init --dry-run --verbose
ai-workspace init --apply
ai-workspace validate

Budget a provider

ai-workspace providers limits openai --mode budgeted --monthly-request-limit 100 --max-output-tokens 2048 --apply

Local-only model workflow

Start Ollama, run providers models ollama, initialize its profile, verify health, and invoke it. No API credential is required.

Safe rollback

Run rollback --dry-run, inspect changed/unmanaged skips, then apply the specific operation with --apply --yes.