LLM Wiki Tools turns your PDFs, documents, notes, and web articles into a living, cross-referenced wiki that Claude builds and maintains for you via MCP.
What is LLM Wiki Tools?
LLM Wiki Tools is an AI-powered knowledge base that transforms uploaded documents into a durable, cross-referenced wiki. It accepts PDFs (with OCR), Word documents, PowerPoint decks, Markdown files, plain text, and web articles as raw sources. Claude, connected via the Model Context Protocol (MCP), reads these sources and writes Markdown wiki pages — including overviews, entity pages, concept pages, summaries, comparison tables, diagrams, logs, and cited notes. Raw sources remain immutable; the wiki is a separate, revisable layer. The product is offered by LLM Wiki Tools (llmwiki.tools).
Key Features
- Compounding knowledge — Every source you add and every question you ask makes the wiki richer. Answers are synthesized once and filed back as pages, never re-derived.
- MCP-native — Connect Claude.ai via a custom MCP server. Claude gains tools to search, read, write, and lint across the wiki vault, turning conversation into authorship.
- Any source — Upload PDFs with OCR, Word, PowerPoint, Markdown, plain text, and web captures. Office formats are converted in an isolated sandbox.
- Citations — Every claim links back to the exact page ranges and passages in the source documents. Open a footnote to view the original PDF.
- Source fidelity — Raw sources are immutable: Claude reads them but never rewrites them. The wiki is fully revisable and reversible.
- Lint & maintain — Ask for a health check. Claude finds contradictions between sources, orphan pages, missing cross-references, and suggests what to read next.
Who is it for?
- Researchers — Upload a collection of PDFs and papers; Claude builds entity pages for each method, cross-references them, and flags contradictions like the one between Chen et al. (2024) and Wei et al. (2023) on emergent capability thresholds.
- Students and lifelong learners — Drop in lecture notes, textbooks, and articles. The wiki grows with each query, turning scattered notes into an organized knowledge base.
- Writers and document-heavy professionals — Curate sources and ask Claude to "write me an entity page on Mistral with citations" or "lint the wiki — what contradicts what?" The wiki compiles itself with cross-references and footnotes.
How does LLM Wiki Tools work?
Follow four steps: (1) Create a wiki at LLM Wiki Tools. (2) Upload your sources — PDFs, Office docs, Markdown, or web captures. (3) Connect Claude via MCP by copying the connector config from Settings into Claude.ai. (4) Tell Claude what to do, e.g., "Read these papers and build an entity page for each method." Claude then searches, reads, writes, and lints the wiki automatically.
FAQ
How is this different from RAG or ChatGPT file upload?
RAG retrieves fragments on each query and forgets them. ChatGPT with files holds context for a single session. An LLM Wiki writes a durable layer — summaries, concept pages, contradictions — that survives across sessions and compounds.
Which LLM does LLM Wiki use?
It connects to Claude via MCP. You use Claude.ai as the interface; the MCP server gives it tools to search, read, write, and delete inside your vault.
Do I need to run anything locally?
No. Sign up at LLM Wiki Tools, upload sources, paste one MCP config into Claude.ai, and you are finished. There is no local setup required.
Who is this based on?
LLM Wiki Tools is a faithful implementation of Andrej Karpathy's LLM Wiki proposal — the three-layer model of raw sources, the LLM wiki, and the schema, with ingest / query / lint as the core operations.
