LLM Wiki
Show HN: 10x better performance from the Coding Harnesses with LLM-wiki
Last verified:
What is LLM Wiki?
LLM Wiki is an open-source implementation of Karpathy’s LLM Wiki pattern that turns raw sources into a structured, continuously maintained knowledge base. It is built around a three-layer model: immutable raw sources, an LLM-generated wiki, and a schema/config layer that controls structure, conventions, and workflows.
The system is designed to ingest documents such as articles, papers, notes, and transcripts, then generate summary pages, entity pages, concept pages, and cross-references. It also flags contradictions between sources, helping the wiki stay internally consistent as new material is added.
A core idea is that the user does not manually write most of the wiki content. Instead, the user curates sources and asks questions, while the LLM handles synthesis, linking, maintenance, and bookkeeping tasks that are tedious to do by hand.
It is aimed at people or teams building serious knowledge bases, especially for research, technical domains, and long-term topic tracking. The website specifically positions it as a way to create an “incredible product” rather than a scattered set of scripts, and it mentions a free, open-source implementation that can be started with raw sources and expanded over time.
LLM Wiki pricing
Pricing model: Freemium
The website presents LLM Wiki as free and open-source, with a “Get started free” message. It does not list paid tiers, subscription plans, or usage-based pricing on the page shown. The product is positioned as something you can start with raw sources and build into a structured wiki without mentioning feature-gated plans or premium editions.
LLM Wiki pros
- Open-source implementation
- Free to get started
- Turns raw sources into a structured wiki
- Maintains the wiki over time
- Supports cross-referenced pages
- Creates summary pages automatically
- Creates entity pages automatically
- Creates concept pages automatically
- Flags contradictions across sources
- Uses immutable raw sources as truth
- Separates content, wiki, and schema cleanly
- Reduces manual bookkeeping
- Handles multiple files in one pass
- Good for complex research questions
- Knowledge compounds over time
- Supports health checks and linting
- Designed for long-term knowledge management
- Works with articles, papers, notes, and transcripts
LLM Wiki cons
- Requires a source-driven workflow
- Not meant for manual wiki writing
- Depends on LLM output quality
- May need schema customization
- Can surface contradictions that still need review
- Best suited to structured knowledge work
- Ingesting sources may create many linked pages to manage
- Value depends on having good raw sources
Frequently asked questions about LLM Wiki
What is LLM Wiki?
LLM Wiki is an open-source system that compiles a structured wiki from raw sources using an LLM. It ingests material like articles, papers, notes, and transcripts, then generates and maintains linked wiki pages from that input.
How does the three-layer model work?
The three layers are raw sources, the wiki, and the schema. Raw sources are the immutable input material, the wiki is the LLM-generated markdown knowledge base, and the schema is the configuration that tells the LLM how to structure pages and run workflows.
Do I write the wiki pages myself?
Not usually. The website says you rarely write the wiki yourself because the LLM owns the wiki layer, while you mainly curate sources and ask questions.
What happens during ingest?
When you drop a source into raw/, the LLM reads it, writes a summary, updates related entity and concept pages, and flags contradictions against existing knowledge. The site says a single source can affect many wiki pages.
What kinds of questions can I ask?
You can ask complex questions against the compiled wiki. The system answers from synthesized knowledge rather than re-deriving every answer from raw chunks each time, so it is meant for deeper comparison and exploration.
What is the lint or health check step?
Linting runs health checks over the wiki to find inconsistent data, stale claims, orphan pages, and missing cross-references. It also suggests new questions to ask and new sources to look for.
Who is LLM Wiki for?
It is aimed at people or teams managing serious knowledge bases, especially in research, technical documentation, and domain tracking. The design is especially useful when the information needs to stay linked, updated, and internally consistent over time.
What content can I put into raw/?
The website lists articles, papers, notes, and transcripts as suitable source material. The raw layer is treated as the source of truth and is not modified by the LLM.
Why is it different from a normal wiki?
Unlike a normal wiki, LLM Wiki emphasizes automated synthesis, cross-linking, contradiction detection, and ongoing maintenance by the LLM. The goal is to make the wiki behave like a dynamic knowledge system instead of a manually maintained document collection.
Is there a free version available?
Yes. The website describes it as a free, open-source implementation and invites users to get started free. No paid plans are shown on the page content provided.