openkb

OpenKB (Open Knowledge Base) is an open-source system that transforms raw documents into a persistent, interlinked wiki-style knowledge base using LLMs. It is powered by PageIndex's vectorless, reasoning-based retrieval, which handles long documents efficiently without a traditional vector database. The tool supports a wide range of formats including PDF, Word, Markdown, PowerPoint, HTML, Excel, CSV, text, and URLs. It automatically generates summaries, concept pages, entity pages (people, orgs, places, products), and cross-references, keeping the knowledge base in sync as new documents are added. OpenKB offers query and chat interfaces for interacting with the compiled wiki, a Skill Factory to distill portable agent skills, and a bundled web UI (Knowledge Workbench) for browsing and managing the knowledge base. It is compatible with Obsidian for graph visualization and integrates with agent CLIs like Claude Code, Codex, and Gemini. The project is Apache 2.0 licensed and built on PageIndex, markitdown, OpenAI Agents SDK, LiteLLM, Click, and watchdog.