Local Deep Research

Local Deep Research is a self-hosted AI research assistant for people investigating complex questions across web sources, academic papers and their own documents. It gathers evidence and writes reports with citations.

• Research depth: Choose a quick research pipeline for factual questions or an autonomous mode that decides which searches to pursue and when to assemble the report.
• Academic and web search: Search arXiv, PubMed and Semantic Scholar alongside web sources, with the autonomous mode choosing engines as the research develops.
• Private knowledge base: Build searchable document collections and research material in your local library, with a private-source mode that restricts searches to local sources and normally refuses cloud inference.
• Model choice: Use local LLMs through Ollama or llama.cpp, connect an OpenAI-compatible endpoint, or use cloud providers such as Google, Anthropic and OpenAI.

The project is open source under the MIT license and supports Docker deployment. Each user has a separate encrypted SQLCipher database that also stores their model API keys. The app has no telemetry, analytics or tracking.

Inference stays on your hardware when the model actually runs there. Web searches need internet access and send queries to search services, including through self-hosted SearXNG; cloud models receive requests at the selected provider. Traffic restrictions are best-effort safeguards. A local endpoint that relays to a hosted model can still send prompts off the machine without the app detecting it.