llmwiki is a knowledge compiler that transforms raw sources such as URLs, files, research papers, documentation sites, notes, and session exports into a structured, interlinked markdown wiki. Unlike retrieval-augmented generation approaches, which retrieve document chunks for each query, llmwiki compiles sources once into a persistent artifact where concepts are organized into typed pages, interlinked as a navigable graph, and every claim is citation-traced back to the original source lines. The tool features a two-phase LLM pipeline that extracts and merges concepts across sources, generates structured wiki pages, and builds chunk-level embeddings for semantic search. Its search functionality narrows results using cosine similarity, reranks with BM25, and expands along the wiki graph to provide citation-traceable evidence packs. Users can browse their compiled wikis through a local web viewer offering sidebar navigation, full-text search, a force-directed page graph, and provenance indicators on each paragraph. llmwiki supports configurable domain profiles through Configurable Lifecycle Profiles (CLP), allowing users to define entities, typed relations, lifecycle state machines, evidence gates, workflows, artifacts, connector bindings, content tiers, and retrieval policies. Built-in workflows such as AutoSci for research and Newsroom for editorial use are available. Projects without a custom profile use a default concepts-and-queries model. The tool also integrates with Google Cloud’s Open Knowledge Format (OKF), enabling export and import of OKF bundles while preserving metadata and provenance. External OKF bundles can be staged for review before being incorporated into the main wiki. Programmatic access is provided via an SDK, allowing in-process control of the pipeline, including OKF import and export. llmwiki can be connected to AI agents through MCP integration, supporting agents like Claude Desktop, Cursor, and Claude Code. Agents can ingest sources, compile, query, lint, retrieve context packs, and exchange OKF bundles without using the command-line interface. The platform is intended for AI researchers, engineers, technical writers, open-source maintainers, and developers seeking to build durable, structured, citation-traceable knowledge bases or to provide persistent context for AI agents. Installation requires Node.js version 24 or higher and an LLM provider credential such as Anthropic, OpenAI-compatible endpoints, Ollama, GitHub Copilot, or local Claude Code login.
llmwiki is a Frameworks & SDKs product. It focuses on automating the compilation of raw sources into a structured, searchable knowledge wiki for persistent reference. llmwiki is an open-source project aimed at developers and researchers managing large knowledge bases. The project is open source (MIT). The product ships for the web, the command line, and API.
It is developed by atomicstrata, and the product first shipped in 2026. The project is developed in the open on GitHub with 1.7k stars and 121 commits in the last 90 days. Across PulseGate's embedding index, llmwiki has few near neighbours, marking it as relatively distinct. Among its 6 catalogued features are semantic search, citation tracing, and markdown export. It exposes integrations via a public API and an MCP server.
Latest indexed changes and source events
atomicstrata/llm-wiki-compiler discovered by the PulseGate indexer
Other apps tracked under the same category.
