llm-wiki-newsroom
Harness engineering applied to knowledge production: a self-evolving multi-agent newsroom that turns your documents into a cross-linked markdown wiki. A "regrou
What is llm-wiki-newsroom?
Harness engineering applied to knowledge production: a self-evolving multi-agent newsroom that turns your documents into a cross-linked markdown wiki. A "reground" loop pulls published pages back in before they go stale — writer ≠ reviewer, local-first, a structured alternative to RAG.
Overview of llm-wiki-newsroom
What it is
LLM Wiki Newsroom is a multi-agent AI knowledge base that turns documents into a cross-linked markdown wiki. It is open-source, local-first, and has no vendor lock-in. The system uses five roles modeled on a newspaper staff: Reporter, Columnist, Copy Editor, Desk, and Editor-in-Chief. The agent that writes a page is never the one that reviews it. The project ships with an example corpus about the debate over what “open source” means for AI, but the framework is domain-agnostic.
Key features
- Five-role newsroom with separation of writing and reviewing - Four loops: Inner (self-check while drafting), Outer (deterministic lint and qualitative review), Meta (authoring guidelines evolve themselves), Reground (published pages come back as input)
- GROUND Ladder governs how much the writer reads before drafting - Authoring guidelines that evolve over time via the meta loop - Interactive knowledge graph with nodes, edges, color-coded clusters, and live physics layout - Contradiction tracking and cascading updates across pages - Output in plain markdown, compatible with Obsidian
Use cases
- Turning collected documents, articles, and PDFs into a structured, cross-referenced wiki - Building a persistent knowledge base as an alternative to RAG - Organizing research material in any domain
Getting started
git clone https://github.com/alfadur7/llm-wiki-newsroom.git cd llm-wiki-newsroom See the repository README for further installation and usage instructions.
License and pricing
The project is licensed under the MIT License and is open source.
Based on the project's public README and repository metadata.