LLM Wiki

LLM Wiki is a desktop AI knowledge base for people building a research library or personal reference collection. It creates lasting, linked pages from your documents and updates them as your sources change.

• Document import: Bring in PDFs, Office files, EPUB/MOBI books, Org files and web clips, with generated pages linked to their original sources.
• Cited answers: Chat with your collection, inspect the pages behind an answer, or restrict answers to original imported material; optional semantic search supports OpenAI-compatible embedding endpoints.
• Knowledge graph: Explore related pages and topic clusters, and find isolated notes or weakly connected areas that need more research.
• Web research: Deep Research searches through Tavily, SerpApi or SearXNG and adds its findings to the wiki; this capability needs access to the selected search service.
• Agent access: The chat agent can search your collection and create workspace files, while a local API and MCP server let other AI tools access wiki content.

The app runs on macOS, Windows and Linux. Its wiki files work as an Obsidian vault, and project archives let you move collections between devices. You can choose separate models for chat and document processing, including models served through Ollama.

PDF extraction can run locally with the built-in parser or local MinerU processing. Choosing MinerU Cloud sends PDFs to that service. Embedded PDF images can receive captions through a vision model, so where that processing happens depends on the model provider you choose.