OpenKB
OpenKB: Open LLM Knowledge Base
OpenKB is an agent skill and CLI system that transforms documents into a structured, interlinked wiki-style knowledge base using LLMs, then packages that knowledge as a portable agent skill. Unlike traditional RAG systems that rediscover information on every query, OpenKB compiles documents once into persistent wiki pages with summaries, concept pages, entity extraction, and cross-references. The resulting knowledge base can be queried directly via CLI, browsed in Obsidian, or installed as a skill in Claude Code, Codex, and Gemini for agent-powered knowledge retrieval. It handles PDFs, Word documents, PowerPoint, Excel, URLs, and more, using PageIndex's vectorless tree indexing for long documents.
Key Features
Use Cases
- 01Building a searchable knowledge base from research papers and technical documentation
- 02Creating domain-expert agent skills from books and specialized materials for Claude Code or Gemini
- 03Maintaining a company wiki that automatically extracts and links entities across documents
- 04Querying large document collections without context limits or vector database overhead
- 05Generating interactive knowledge graphs and HTML slide decks from compiled wikis
- 06Integrating compiled knowledge into agent workflows via native skill installation
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
OpenKB — FAQ
What is OpenKB?+
OpenKB is an open-source CLI system that compiles documents into a structured wiki-style knowledge base using LLMs, then packages it as an agent skill that Claude Code, Codex, and Gemini can install and query. It uses reasoning-based retrieval via PageIndex instead of vector databases.
How do I install OpenKB?+
Install via pip with 'pip install openkb', then run 'openkb init' in a new directory to initialize a knowledge base. For the web UI, use 'pip install openkb[web]' and start with 'openkb-web'.
Which AI clients support OpenKB skills?+
OpenKB generates agent skills compatible with Claude Code, OpenAI Codex CLI, and Gemini CLI. The skill uses the SKILL.md format and can be installed via each client's native skill marketplace or manual installation.
Do I need API keys to use OpenKB?+
Yes, you need an LLM API key (OpenAI, Anthropic Claude, Google Gemini, etc.) set in a .env file as LLM_API_KEY. PageIndex runs locally by default, but an optional PAGEINDEX_API_KEY enables cloud features like OCR for scanned PDFs.
Is OpenKB free and open source?+
Yes, OpenKB is open source under the Apache 2.0 license. You will incur LLM API costs from your chosen provider (OpenAI, Anthropic, etc.) based on usage.
How does OpenKB handle long documents?+
Documents under 20 pages are read directly by the LLM. Longer PDFs are processed by PageIndex into a hierarchical tree index that enables accurate, context-aware retrieval without hitting context limits or using vector databases.
How do I install OpenKB?+
Open the source repository on GitHub and follow its README. OpenKB is a skill — MCP Agents Market links you directly to the official repo.
Is OpenKB free?+
OpenKB is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.