AI transparency
How AI features use data
AI features are optional, scoped, and hidden when no model provider is configured.
What may be sent
Only content needed for the requested feature: selected lesson or roadmap context, entered text, selected notes, or a size-capped repository snapshot.
What is not sent automatically
Authentication tokens, encrypted GitHub credentials, unrelated notes, and content outside the selected scope are not added to prompts.
Storage and control
Generated artifacts, chat messages, usage counts, and review results may be stored so the feature works and costs can be audited. Account deletion removes user-owned records subject to the privacy policy.
Human verification
AI progress proposals never save automatically. Generated content and findings can be wrong; verify important output and never submit secrets.