LLM Wiki: Your Self-Building, Interlinked Knowledge Base
Transform your documents into an intelligently organized and interconnected personal wiki with LLM Wiki. This powerful cross-platform desktop application goes beyond traditional RAG by incrementally building and maintaining a persistent knowledge base, ensuring your information is always current and readily accessible.
Key Features & Innovations:
- Self-Organizing Knowledge: LLM Wiki automatically structures your imported documents into a dynamic wiki, creating interlinked pages for entities, concepts, and more.
- Advanced Ingest Process: Experience a two-step Chain-of-Thought ingest for superior analysis and generation, with intelligent incremental caching to save time and resources. Supports a wide array of document formats, including PDFs, Office docs, EPUB, and images.
- Multimodal Intelligence: Ingest embedded images from PDFs, generate factual captions with vision LLMs, and find them with image-aware search.
- Dynamic Knowledge Graph: Visualize and analyze your knowledge with a 4-signal relevance model and Louvain community detection, uncovering surprising connections and knowledge gaps.
- Intelligent Querying: Benefit from a multi-phase retrieval pipeline, including optional vector semantic search, for accurate and comprehensive answers.
- Persistent & Powerful Chat: Engage in multi-conversation chat sessions with a Rust backend agent capable of tool usage, custom skill integration, and safe execution of workspace commands.
- Deep Research Engine: Automatically identify knowledge gaps and trigger LLM-optimized web research, synthesizing findings directly into your wiki.
- Seamless Workflow: Enjoy features like folder auto-watching, a persistent ingest queue, and a Chrome web clipper for effortless knowledge capture.
- Obsidian Compatibility: Your LLM Wiki works directly as an Obsidian vault, offering a familiar and flexible user experience.
LLM Wiki represents a significant leap forward, building upon Andrej Karpathy's foundational LLM Wiki pattern and delivering a robust, user-friendly desktop application with extensive enhancements.
