DocsetMCP

Access your local Dash documentation directly from AI assistants 🚀

DocsetMCP is a Model Context Protocol (MCP) server that seamlessly integrates your local Dash docsets with AI assistants like Claude, enabling instant access to offline documentation without leaving your conversation.

📋 Table of Contents

• Why DocsetMCP?

• Quick Start

• Features

• Prerequisites

• Installation

• Configuration

• Usage Examples

• Available Tools

• Troubleshooting

• Development

• Contributing

• License

Why DocsetMCP?

• 📚 Instant Documentation: No switching, no web searches. Get straight to the docs directly in your AI conversation

• 🔒 Local and Private: Work with docset files on your machine

• ⚡ Lightning Fast: Optimized caching and direct database queries

• 🎯 Precise Results: Get exactly what you need with smart filtering

Quick Start

Add to your MCP config and restart your MCP client. Then try asking something like "Find me the AppIntent documentation"

✨ Features

Documentation Search

• Multi-Docset Support: Search across 165+ supported docsets including Apple, NodeJS, Python, and more

• Language Filtering: Target specific programming languages within docsets

• Name-Based Search: Only returns entries where search terms match item names for precise results

• Smart Ranking: Results ranked by match type (exact > prefix > substring) and dynamic type ordering

• Container Guidance: Framework and class entries show drilldown notes for exploring members

Cheatsheet Access

• Quick Reference: Instant access to Git, Vim, Docker, and 40+ other cheatsheets

• Fuzzy Matching: Find cheatsheets even with partial names

• Category Browsing: Explore commands by category within each cheatsheet

• Search Within: Query specific commands inside any cheatsheet

Performance & Integration

• Efficient Caching: In-memory caching for repeated queries

• Direct Database Access: No intermediate servers or APIs

• Universal: Works with Claude Desktop, Cursor, VS Code, and any MCP-compatible client

• Framework Discovery: List all available frameworks/types in any docset

• Container Guidance: Automatic drilldown notes for frameworks and classes with members

📦 Supported Docsets

DocsetMCP supports 165+ docsets including:

• Python (2 & 3)

• JavaScript / TypeScript

• Java

• C / C++

• Go

• Rust

• Ruby

• Swift / Objective-C

• PHP

• Bash

• And many more...

• React / Angular / Vue

• Node.js / Express

• Django / Flask

• Ruby on Rails

• Bootstrap

• jQuery

• And many more...

• Git (cheatsheet)

• Docker (cheatsheet)

• Vim (cheatsheet)

• MySQL / PostgreSQL

• MongoDB / Redis

• nginx / Apache

• And many more...

Use list_available_docsets to see all docsets installed on your system.

Prerequisites

• macOS (Dash is Mac-only)

• Dash with desired docsets downloaded

• Python 3.10 or higher

• UV package manager (How to Install)

• An AI assistant that supports MCP (Claude Desktop, Claude Code CLI, Cursor IDE, etc.)

Configuration

Custom Docset Locations

By default, DocsetMCP looks for docsets in Dash's standard directories:

• Docsets: ~/Library/Application Support/Dash/DocSets

• Cheatsheets: ~/Library/Application Support/Dash/Cheat Sheets

You can customize these locations using:

Environment Variables

Command Line Arguments

Priority Order:

1. CLI arguments (highest priority)

2. Environment variables

3. Default Dash locations (lowest priority)

Additional Search Paths:

The --additional-docset-paths and --additional-cheatsheet-paths options allow DocsetMCP to search in multiple locations beyond the primary path. This is useful when:

• You have docsets in multiple directories

• You want to include third-party or custom docsets

• You're sharing docsets across different tools

DocsetMCP will automatically discover and configure docsets found in these additional paths.

MCP Client Setup

Choose your MCP client below for specific setup instructions:

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

For custom docset locations:

Add to your MCP configuration (Cursor: .mcp/mcp.json in your project root:

Note: Restart your client and check your MCP settings for connection status.

Installation

No Installation Required (Recommended)

If your MCP client supports uvx, no installation is needed! The package will be automatically downloaded and run when needed. See the Quick Start or Configuration sections.

Manual Installation

If you prefer to install locally or your MCP client doesn't support uvx:

Then use docsetmcp instead of uvx docsetmcp in your configuration.

Development Installation

1. Clone and install:

   git clone https://github.com/codybrom/docsetmcp.git cd docsetmcp pip install -e .

2. Run tests (optional):

   # Install test dependencies pip install pytest pytest-cov pytest-xdist # Run basic tests pytest tests/test_docsets.py::TestDocsets::test_yaml_structure -v # Run quick tests (structure + existence checks) pytest tests/ -k "yaml_structure or test_docset_exists" -v # Run full test suite (all docsets) pytest tests/ -v # Run with coverage pytest tests/ --cov=docsetmcp --cov-report=html -v # Validate all local cheatsheets work (integration test) python scripts/validate_cheatsheets.py

Usage Examples

Once configured, you can ask your AI assistant to search documentation naturally:

🍎 iOS/macOS Development

🌐 Web Development

🛠️ DevOps & Terminal

📊 Data Science

Advanced Usage

Discovery Workflow

DocsetMCP is designed for name-based searches, not keyword searching. Follow this workflow:

1. Start with Discovery Tools

2. Then Search by Exact Names

3. Use Drilldown Notes

When you find container types (frameworks, classes), follow the drilldown guidance:

How It Works

1. Multi-Format Support: Handles both Apple cache format and tarix compression

2. Direct Database Access: Queries Dash's SQLite databases for fast lookups

3. Name-Based Matching: Only returns entries where search terms match item names (no false positives)

4. Smart Ranking: Prioritizes exact matches, then prefix matches, then substring matches

5. Dynamic Type Ordering: Uses docset configuration files for intelligent result prioritization

6. Container Detection: Automatically detects frameworks/classes with members and provides exploration guidance

7. Smart Extraction: Decompresses Apple's DocC JSON or extracts HTML from tarix archives

8. Markdown Formatting: Converts documentation to readable Markdown

Available Tools

DocsetMCP provides eleven powerful tools for accessing your documentation:

🔍 search_docs

Search and extract documentation from any docset.

📋 search_cheatsheet

Search Dash cheatsheets for quick command reference.

📚 list_available_docsets

List all installed Dash docsets with their supported languages.

📝 list_available_cheatsheets

List all available Dash cheatsheets that can be searched.

🏗️ list_frameworks

List frameworks/types within a specific docset.

🌍 list_languages

Discover all programming languages with available documentation.

📖 list_docsets_by_language

Find all docsets that support a specific programming language.

🏷️ list_types

List all available types (Class, Protocol, Function, etc.) in a docset/language.

📋 list_entries

List entries filtered by type and optional name prefix.

📂 list_cheatsheet_categories

List all categories within a specific cheatsheet.

📄 fetch_cheatsheet

Fetch entire cheatsheet content (recommended for comprehensive access).

Troubleshooting

This means the docset isn't installed in Dash. To fix:

1. Open Dash.app

2. Go to Preferences → Downloads

3. Download the required docset

4. Restart your MCP client

1. Check installation: Run pip show docsetmcp to verify installation

2. Test manually: Run uvx docsetmcp in terminal - you should see MCP output

3. Check logs:

   • Claude Desktop: Check Console.app for Claude logs

   • Cursor: Check Output → MCP panel

4. Verify config path: Ensure config file is in the correct location

• The content might not be in your local Dash cache

• Try searching with different terms or partial matches

• Use list_available_docsets to verify the docset is loaded

• Some docsets may use different naming conventions (e.g., 'fs' vs 'filesystem')

1. Python version: Ensure you have Python 3.10 or higher

2. UV not found: Install UV package manager from https://docs.astral.sh/uv/

3. Permission denied: Check file permissions on Dash docsets directory

4. Report bugs: Open an issue at https://github.com/codybrom/docsetmcp/issues

Development

Building from Source

Testing

Code Quality

CLI Commands

Building Distribution

Architecture

Core Components

• docsetmcp/server.py: Main MCP server implementation using FastMCP. Contains the DashExtractor class that handles:

  • Apple cache format (SHA-1 UUID-based with brotli compression)

  • Tarix format (tar.gz archives)

  • SQLite database queries for documentation lookup

  • HTML to Markdown conversion

• docsetmcp/config_loader.py: Configuration system that loads YAML configs for 165+ supported docsets. Provides smart defaults and handles both simple and complex configuration formats.

• docsetmcp/docsets/: YAML configuration files for each supported docset, defining:

  • Docset paths and formats

  • Language variants and filters

  • Type priorities for search results

Key Implementation Details

1. Multi-Format Support: The server detects and handles both Apple's modern cache format (using SHA-1 based UUIDs) and the older tarix compression format automatically based on docset configuration.

2. Caching Strategy: Extracted documentation is cached in memory (_fs_cache for Apple format, _html_cache for tarix) to improve performance on repeated queries.

3. Search Algorithm: Uses SQLite case-insensitive LIKE queries on the optimizedIndex.dsidx database. Results are ranked by match type (exact > prefix > substring) and then by dynamic type ordering from docset configuration files. Only returns entries where the search term matches the item name.

4. Configuration Loading: The ConfigLoader applies smart defaults, allowing minimal YAML configs while supporting complex overrides when needed.

5. Container Type Detection: Framework, class, and module entries automatically include drilldown notes when they contain additional members, guiding users to search for more specific content.

Contributing

We welcome contributions! Here's how you can help:

Adding New Docset Support

1. Create a YAML configuration in docsetmcp/docsets/:

   # docsetmcp/docsets/my_docset.yaml name: My Docset description: Brief description of the docset docset_path: My_Docset/My_Docset.docset languages: - python - javascript

2. Test your configuration:

   pytest tests/test_docsets.py -k "my_docset" -v

3. Submit a pull request

Reporting Issues

• 🐛 Bug Reports

• 💡 Feature Requests

• 📚 Documentation Issues

Development Guidelines

• Follow PEP 8 style guidelines

• Add tests for new features

• Update documentation as needed

• Keep commits focused and descriptive

Technical Architecture

DocsetMCP leverages Dash's internal structure for efficient documentation access:

• Format Support: Handles both Apple's modern cache format (SHA-1 UUID-based with brotli compression) and traditional tarix archives

• Caching Strategy: In-memory caching for repeated queries

• Database Access: Direct SQLite queries to Dash's optimized indexes

• Content Extraction: Smart extraction with fallback strategies

• Type System: Full type hints for better IDE support

License

MIT License - see LICENSE file for details.

Acknowledgments

• Thanks to Kapeli for creating Dash

• Built on the Model Context Protocol standard

• Inspired by the MCP community and ecosystem