AI features overview

Pro AI capability map, entry points, setup prerequisites, and conservative policy

Prerequisites

1. Activate Pro → see the entitlement table in Plan comparison
2. Configure a Provider under Settings → AI (BYOK, step-by-step setup)
3. Pick default models for each task: Chat (summary / chat / tags) and Embedding (semantic search / RAG)

Missing Embedding is the most common reason you have Pro but still cannot use semantic search or RAG.

Capability map

Feature	Typical entry	Docs
AI summary	Detail panel ⇧⌘A	Summary
AI chat	Same panel; can detach to a separate window	Chat
Tag recommendations	Panel / Auto Tidy	Tag recommendations
README translation	README toolbar	Translation
AI personal notes	Generate with AI in the notes area	Personal notes
Semantic search	List semantic mode	Semantic search
RAG Q&A	⇧⌘K	Knowledge base
Auto Tidy	Settings → AI	Auto Tidy
Agent workspace	Separate window	Agent Beta

Notes:

• The web scope in Search Center does not require Pro
• If AI attaches live web context, Pro and toggle rules still apply → External search

Conservative policy

Scenario	Rule
Tag recommendations	Written only after you confirm
AI notes	Saved only after you confirm the draft
Agent	Data changes require explicit authorization
Auto Tidy	Runs only the actions and triggers you enable

Recommended onboarding order

1. Configure models

   AI configuration → test connection.
2. Try one repo

   Detail → ⇧⌘A → summary or ask one question.
3. When you need semantic search

   Add Embedding → build index → list semantic mode.
4. When you need cross-repo Q&A

   Curate ingest → RAG workspace.
5. Then automate

   After output quality looks good, enable Auto Tidy.

Related

• AI settings overview
• FAQ · AI