LLM Wiki: Your Self-Building, Interlinked Knowledge Base
Transform your documents into a dynamic, interconnected wiki that evolves with you. LLM Wiki is a powerful cross-platform desktop application that goes beyond traditional knowledge retrieval. Instead of repeatedly processing your sources, LLM Wiki incrementally builds and maintains a persistent, structured knowledge base, ensuring your information is always current and efficiently organized.
Key Features:
- Automated Wiki Construction: LLM Wiki intelligently parses your documents, creating a web of interlinked knowledge pages.
- Persistent & Incremental Learning: Unlike reactive RAG systems, LLM Wiki learns and updates your knowledge base over time, remembering and building upon previous ingestions.
- Rich Document Ingestion: Supports a wide array of formats including PDFs, Office documents, EPUBs, images, media, and web clips.
- Advanced Knowledge Graph: Visualize and analyze connections between your knowledge with a 4-signal relevance model and Louvain community detection for automatic cluster discovery.
- Intelligent Querying & Deep Research: Experience source-grounded retrieval, semantic vector search, and powerful "Deep Research" capabilities that automatically explore and synthesize information from the web into your wiki.
- Cross-Platform Desktop Application: Enjoy a seamless user experience with a feature-rich interface, customizable panels, and persistent state management.
- Obsidian Compatibility: Your LLM Wiki functions as a fully compatible Obsidian vault, allowing for flexible workflows.
- Multimodal Ingestion: Extracts embedded images from PDFs and generates factual captions for enhanced searchability.
- Flexible Model Configuration: Easily manage and route different LLM models for chat and ingestion tasks.
- Browser Integration: Capture web content directly into your knowledge base with a convenient Chrome Extension.
LLM Wiki is built on the innovative LLM Wiki pattern by Andrej Karpathy, enhanced with a robust desktop application architecture, a sophisticated knowledge graph, and a powerful AI agent for an unparalleled personal knowledge management experience.
