locallens index

Index local files into the vector database for semantic search.

Usage

locallens index <folder_path> [OPTIONS]

Arguments

Argument	Required	Description
folder_path	Yes	Path to the folder to index

Options

Option	Default	Description
--force	false	Re-index all files, ignoring hash cache
--namespace	default	Namespace to index into
--format	rich	Output format: rich or json

Examples

# Index a folder
locallens index ~/Documents

# Force re-index everything
locallens index ~/Documents --force

# Index into a specific namespace
locallens index ~/project --namespace myproject

JSON output

locallens index ~/Documents --format json

{
  "total_files": 42,
  "new_files": 5,
  "updated_files": 2,
  "skipped_files": 35,
  "total_chunks": 1284,
  "duration_seconds": 3.47
}

How it works

1. Scans the folder recursively for supported file types
2. Computes a SHA-256 hash of each file's content
3. Skips files whose hash already exists in the index (unless --force)
4. Extracts text using the appropriate extractor
5. Chunks text into ~500 character segments with 50 character overlap
6. Embeds each chunk using sentence-transformers
7. Upserts into Qdrant with deterministic point IDs