vector-notes-indexer

To build a local semantic search index, run python3 scripts/main.py to generate vector embeddings from your notes and store them on-device using ChromaDB, enabling offline content retrieval without external services.

Yes, offline semantic retrieval supports indexing multiple formats including Markdown, plain text, and PDF files. Vector embeddings are generated for the text content to capture meaning and enable search across your local knowledge base.

Incremental updates let you add or update notes without requiring a full reindexing of the entire collection. The indexer processes only the new or modified content, generating vector embeddings and appending them to your local storage.

Yes, local note indexing requires sentence-transformers to generate vector embeddings and ChromaDB to store the indexes on-device. These dependencies enable offline batch indexing and semantic search capabilities without cloud access.

Yes, you can configure embedding models and index parameters to suit different domains. This customization allows the semantic search to better capture meaning specific to your personal knowledge bases or documentation collections.

The best way to batch index a large documentation collection offline is processing the files locally to generate vector embeddings efficiently. This approach supports multiple formats and stores indexes on-device for offline semantic retrieval.