Search · AI
Beyond Grep: AI-powered semantic search with 'vem ask'
Use vector embeddings and RAG to answer complex architectural questions across your entire project history.
Question-led retrieval
Traditional search is limited by keywords. If you ask 'How does the auth flow work?', grep might miss the conceptual links between the middleware, the database schema, and the client-side hooks.
The new `vem ask` command uses Gemini-powered semantic search and RAG (Retrieval-Augmented Generation) to synthesize an answer from your tasks, decisions, and actual source code.
Ask a project question
vem ask "How is MFA enforced in the web app?"Citations you can trust
An answer is only as good as its sources. `vem ask` provides direct citations to the exact code chunks and memory artifacts it used to build its response.
This lets you jump directly from a high-level summary to the implementation details, saving hours of manual discovery.