FAQ

Do I need Docker?

No. The Python library and CLI use Qdrant Edge, an embedded vector database that runs in-process. Docker is only needed for the web dashboard, which uses a Qdrant server container.

Do I need Ollama?

Only for the ask / RAG feature. Search, index, stats, and doctor all work without Ollama. Install Ollama only when you want to ask natural language questions about your files.

How much RAM does it need?

About 2-3 GB with everything loaded (embedding model + Qdrant Edge shard). 4 GB recommended. The embedding model (all-MiniLM-L6-v2) is small at 80 MB.

Does it work on Windows?

Yes. LocalLens works on Python 3.11+ on Windows, macOS, and Linux. The voice feature (moonshine-voice, piper-tts) works best on macOS and Linux.

Can I use a different LLM?

Yes. Any Ollama model works. Set the model when creating the instance:

lens = LocalLens("~/Documents", ollama_model="llama3.2:3b")

Or via environment variable:

export LOCALLENS_OLLAMA_MODEL=llama3.2:3b

Can I use a different embedding model?

Yes. Any sentence-transformers compatible model works:

lens = LocalLens("~/Documents", embedding_model="BAAI/bge-small-en-v1.5")

WARNING

Changing the embedding model requires re-indexing. Delete ~/.locallens/qdrant_data and re-run locallens index.

How do I update the index when files change?

Re-run locallens index <folder>. It uses content hashing to automatically skip unchanged files — only new or modified files get re-indexed. Use --force to re-index everything.

For automatic re-indexing, use watch mode:

locallens watch ~/Documents

Is my data sent anywhere?

No. Everything runs on your machine:

• Embeddings computed locally via sentence-transformers
• Vector storage in Qdrant Edge (a file on disk)
• LLM inference via local Ollama

Zero network calls. No cloud APIs. No telemetry.

How is this different from other file search tools?

LocalLens combines several capabilities in one offline package:

• Hybrid search (semantic + BM25 with RRF fusion)
• RAG Q&A with source citations
• Voice interface (STT + TTS)
• MCP server for AI agent integration
• Structured JSON output for scripting
• Plugin extractors for custom file formats