xs:workspace-rag

To search your entire workspace, you can use semantic vector search to instantly find specific information scattered across various file types. It utilizes a lightweight SQLite database and the multilingual-e5-small embedding model to retrieve relevant snippets quickly.

Semantic search finds information in local files by converting text into embeddings and storing them in a SQLite database. This allows the system to match meaning rather than exact keywords, returning relevant snippets with relevance scores via an R²AG-inspired approach.

Yes, you need to install PyTorch and sentence-transformers because the skill relies on these dependencies to generate embeddings. These libraries enable the multilingual-e5-small model required for processing your local development environment files.

The best way to keep your workspace search index updated is using differential indexing. This feature quickly updates the SQLite database by processing only the files that have changed, ensuring efficient and fast index refreshes.

Yes, you can use SQLite for vector search across diverse file types in your project. This skill stores embeddings in SQLite and supports differential indexing to rapidly retrieve information from various local development files.