Skip to content

vector-knowledge.md

Latest commit

 

History

History
18 lines (12 loc) · 1017 Bytes

vector-knowledge.md

File metadata and controls

18 lines (12 loc) · 1017 Bytes

Vector knowledge & LlamaIndex helper

Vector search is optional but offered to boost retrieval quality when you want the agent to cite cached embeddings.

Enable vector search

  1. Set vector_search.enabled to true in config.json.
  2. Run npm run vector-bake to build the store (requires Python and llama-index).
  3. The CLI automatically references data/vector-store/ inside prompts when vector_search.enabled is true.

Manage the store

  • npm run vector-bake – Bake docs + GitHub issues/reviews.
  • npm run vector-bake -- --docs-only – Bake only documentation (skip GitHub content).
  • npm run vector-bake -- --query "How do I ...?" – Query the store and print the top similarity hits.
  • npm run vector-bake -- --clean – Remove data/vector-store/ and rebuild from scratch (useful after large doc changes).

The script installs llama-index automatically if missing. Keep Python 3.9+ available (the CLI detects python3/python, then runs the embedded bake script).