llmwiki

Overview

LLM Wiki provides a complete local-first solution for building a knowledge base from your research documents. It indexes files in your directory and exposes MCP tools that allow Claude to read, write, organize, and maintain wiki pages. The system supports multiple document formats including PDF, Markdown, HTML, Excel, and more. All processing runs locally with optional cloud storage for hosting. The wiki pages are stored as ordinary Markdown files that can be edited with any text editor.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• guide

  Explains how the wiki works, lists what's in the workspace

  Try:Explain how this wiki worksTry:What's in this workspace?
• search

  Browse files or full-text search

  Try:Search for documents about attentionTry:List all PDF files in the folder
• read

  Read documents — PDFs with page ranges, glob batch reads

  Try:Read paper.pdf pages 1-5Try:Read all markdown files in the notes folder
• write

  Create wiki pages, edit with str_replace, append. SVG/CSV assets

  Try:Create a new wiki page about transformersTry:Add content to concepts/attention.md
• delete

  Delete documents by path or glob pattern

  Try:Delete the old draft documentTry:Remove all temporary files

Comparable tools

Installation

# Clone and setup
git clone https://github.com/lucasastorian/llmwiki.git
cd llmwiki

cd api && python -m venv .venv && source .venv/bin/activate && pip install -r requirements.txt
cd .. && cd web && npm install && cd ..

# Initialize workspace
./llmwiki init ~/research

# Start servers
./llmwiki serve ~/research

# Generate Claude Desktop config
./llmwiki mcp-config ~/research

{
  "mcpServers": {
    "llmwiki-research": {
      "command": "python",
      "args": ["-m", "llmwiki", "mcp", "~/research"]
    }
  }
}

FAQ

What document formats are supported?

LLM Wiki supports PDF, Markdown/Text, HTML, Excel/CSV, Images, and Word/PowerPoint (with LibreOffice). PDF processing uses pdf-oxide by default with optional Mistral API for better OCR.

Can I use this with multiple research projects?

Yes, each research folder runs as a separate MCP server. You'll need to add one MCP entry per folder in your Claude Desktop configuration.

Compare llmwiki with