Labsco
Quinncuatro logo

Jane

โ˜… 3

from Quinncuatro

A knowledge management server for stdlib and specs documents, with a configurable storage path.

๐Ÿ”ฅ๐Ÿ”ฅ๐Ÿ”ฅโœ“ VerifiedFreeQuick setup

Jane - Knowledge Management for AI Teams

Jane is a Model Context Protocol (MCP) server that transforms your team's documentation into an AI-accessible knowledge base. Built with TypeScript and the official MCP SDK, Jane allows Claude and other MCP-compatible clients to search, access, and manage your standard library documentation and project specifications with SQLite-powered performance.

Perfect for development teams who want to:

  • ๐Ÿ“š Make their documentation instantly searchable by AI assistants
  • ๐Ÿ” Quickly find APIs, coding standards, and architectural decisions
  • ๐Ÿ“ Keep project specifications and requirements accessible
  • ๐Ÿš€ Onboard new team members faster with searchable knowledge

Key Features

  • ๐Ÿ” Smart Document Search: SQLite-powered full-text search with intelligent indexing that only processes new/modified documents
  • ๐Ÿค– AI-Native Integration: Seamlessly works with Claude Desktop, Claude Code, and custom MCP clients
  • ๐Ÿ“Š Performance Optimized: Context window preservation through efficient SQLite database storage
  • ๐Ÿ“ Flexible Organization: Supports both standard library docs and project specifications
  • โšก Zero Configuration: Works out of the box with sensible defaults and automatic setup

MCP Tools

Jane provides 7 MCP tools for comprehensive document management:

Discovery Tools

  • list_stdlibs - Browse available programming languages and their documentation
  • list_specs - Explore your project specifications and technical docs
  • search - Full-text search across all documents with filtering options

Document Access

  • get_stdlib - Retrieve specific standard library documentation
  • get_spec - Access project specifications and technical documents

Content Management

  • create_document - Add new documentation with structured metadata
  • update_document - Modify existing documents and their metadata

Integration Guides

Claude Desktop Setup

  1. Build Jane (if not done already):

    npm run build
  2. Find your absolute path:

    pwd
    # Copy the output, you'll need: /that/path/dist/index.js
  3. Update Claude Desktop config with the JSON above

  4. Restart Claude Desktop completely (quit and reopen)

  5. Test the integration:

    • Look for the ๐Ÿ“Ž icon in the input area
    • Select "jane" from the MCP servers dropdown
    • Try: "List available standard libraries"

Claude Code Integration

Add Jane as an MCP server to Claude Code:

# Add Jane MCP server to Claude Code
claude mcp add jane node /absolute/path/to/jane-mcp-server/dist/index.js

# Start Claude Code interactive session
claude

Verify the connection:

# List configured MCP servers
claude mcp list

# Test Jane integration
> /mcp
# Select "jane" from the interactive menu to verify connection

For teams who want to keep their documentation private and separate from the MCP server repository:

Setup Private Document Storage

# Create your private Jane documents directory
mkdir -p ~/Documents/Jane/{stdlib,specs}

# Remove default Jane directory (if it exists)
rm -rf ~/dev/jane-mcp-server/Jane

# Create symlink from MCP server to your private docs
ln -s ~/Documents/Jane ~/dev/jane-mcp-server/Jane

Benefits

  • Privacy: Keep your docs separate from the open-source MCP server
  • Backup: Back up docs independently from the server code
  • Team Sharing: Share docs privately while using the same server setup
  • Git Cleanliness: Avoid committing sensitive documentation to the MCP server repo

Usage

After setting up the symlink, Jane will automatically use your private documents from ~/Documents/Jane while the server accesses them through the symlink. You can add .gitignore entry for Jane/ to prevent accidentally committing your private docs.

Development

# Development mode with hot reload
npm run dev

# Run tests
npm test

# Lint code
npm run lint

# Check SQLite database
sqlite3 document-index.db "SELECT COUNT(*) FROM documents;"