Fabric AI MCP Server

A Model Context Protocol (MCP) server that integrates Fabric AI framework with AI assistants.

Built with TypeScript for type safety and modern development practices.

Purpose

This MCP server provides a secure interface for AI assistants to use Fabric AI's 234+ patterns for augmenting human capabilities with AI. Fabric is an open-source framework that provides modular AI prompts for solving specific problems.

Features

Current Implementation

• execute_pattern - Execute any Fabric pattern on input text
• list_patterns - List all 234+ available patterns with optional search
• get_pattern_details - Get information about a specific pattern
• process_url - Fetch and process web content with a pattern
• process_youtube - Extract YouTube transcripts and process with a pattern
• update_patterns - Update to the latest Fabric patterns

Popular Patterns

• extract_wisdom - Extract insights and wisdom from content
• summarize - Create comprehensive summaries
• improve_writing - Enhance text quality
• explain_code - Explain code clearly
• analyze_presentation - Analyze presentation quality
• create_summary - Create structured summaries
• youtube_summary - Summarize YouTube videos

Prerequisites

• Node.js 20 or higher
• Docker Desktop with MCP Toolkit enabled
• Docker MCP CLI plugin (docker mcp command)
• Fabric AI CLI installed (installation handled in Docker)

Installation

Step 1: Clone and Install Dependencies

git clone https://github.com/gufao/mcp-server-fabric-ai.git
cd mcp-server-fabric-ai
npm install

Step 2: Build TypeScript

npm run build

Step 3: Build Docker Image

docker build -t fabric-ai-mcp-server .

Step 4: Create Custom Catalog

# Create catalogs directory if it doesn't exist
mkdir -p ~/.docker/mcp/catalogs

# Create or edit custom.yaml
nano ~/.docker/mcp/catalogs/custom.yaml

Add this entry to custom.yaml:

version: 2
name: custom
displayName: Custom MCP Servers
registry:
  fabric-ai:
    description: "Fabric AI framework integration - 234+ patterns for AI-augmented workflows"
    title: "Fabric AI"
    type: server
    dateAdded: "2025-11-21T00:00:00Z"
    image: fabric-ai-mcp-server:latest
    ref: ""
    readme: ""
    toolsUrl: ""
    source: ""
    upstream: "https://github.com/danielmiessler/fabric"
    icon: ""
    tools:
      - name: execute_pattern
      - name: list_patterns
      - name: get_pattern_details
      - name: process_url
      - name: process_youtube
      - name: update_patterns
    secrets: []
    metadata:
      category: productivity
      tags:
        - ai
        - patterns
        - automation
        - content-processing
      license: GPL-3.0
      owner: local

Step 5: Update Registry

nano ~/.docker/mcp/registry.yaml

Add this entry under the existing registry: key:

registry:
  # ... existing servers ...
  fabric-ai:
    ref: ""

IMPORTANT: The entry must be under the registry: key, not at the root level.

Step 6: Configure Claude Desktop

Find your Claude Desktop config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

Edit the file and add your custom catalog to the args array:

{
  "mcpServers": {
    "mcp-toolkit-gateway": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-v", "/var/run/docker.sock:/var/run/docker.sock",
        "-v", "[YOUR_HOME]/.docker/mcp:/mcp",
        "docker/mcp-gateway",
        "--catalog=/mcp/catalogs/docker-mcp.yaml",
        "--catalog=/mcp/catalogs/custom.yaml",
        "--config=/mcp/config.yaml",
        "--registry=/mcp/registry.yaml",
        "--tools-config=/mcp/tools.yaml",
        "--transport=stdio"
      ]
    }
  }
}

Replace [YOUR_HOME] with:

• macOS: /Users/your_username
• Windows: C:\\Users\\your_username (use double backslashes)
• Linux: /home/your_username

Step 7: Restart Claude Desktop

1. Quit Claude Desktop completely
2. Start Claude Desktop again
3. Your new tools should appear!

Step 8: Test Your Server

# Verify it appears in the list
docker mcp server list

Development

Local Development

# Install dependencies
npm install

# Run in development mode with auto-reload
npm run dev

# Type check
npm run typecheck

# Build
npm run build

# Run production build
npm start

Local Testing

# Install Fabric CLI locally first
go install github.com/danielmiessler/fabric/cmd/fabric@latest

# Set up Fabric (configure API keys)
fabric --setup

# Run the server
npm start

Usage Examples

In Claude Desktop, you can ask:

• "Use Fabric to extract wisdom from this article: [paste text]"
• "Summarize this YouTube video using Fabric: https://youtube.com/watch?v=..."
• "List all available Fabric patterns related to writing"
• "Use the improve_writing pattern on this draft: [paste text]"
• "Analyze this presentation using Fabric patterns"
• "Get details about the extract_wisdom pattern"

Architecture

Claude Desktop -> MCP Gateway -> Fabric AI MCP Server -> Fabric CLI -> AI Models
                       |
            Docker Desktop
            (Fabric installed)

TypeScript Benefits

• Type Safety: Catch errors at compile time
• Better IDE Support: Enhanced autocomplete and refactoring
• Modern JavaScript: Use latest ECMAScript features
• Maintainability: Self-documenting code with types

Adding New Tools

To extend functionality:

1. Add tool function in src/index.ts:

async function myNewTool(param: string): Promise<string> {
  // Implementation
  const result = await execAsync(`fabric --pattern ${param}`);
  return `Result: ${result.stdout}`;
}

2. Add tool definition to TOOLS array
3. Add case to tool handler switch statement
4. Rebuild: npm run build && docker build -t fabric-ai-mcp-server .

Troubleshooting

Tools Not Appearing

• Verify Docker image built successfully
• Check catalog and registry files
• Ensure Claude Desktop config includes custom catalog
• Restart Claude Desktop

Build Errors

• Check TypeScript version compatibility
• Run npm run typecheck to see type errors
• Ensure all dependencies are installed

Fabric Not Found

• Verify Fabric CLI is installed in the Docker container
• Check Dockerfile installation steps
• Ensure fabric is in PATH

Pattern Execution Errors

• Verify pattern name is correct: fabric --listpatterns
• Check if Fabric is properly configured: fabric --setup
• Ensure API keys are set for AI models

Security Considerations

• Fabric AI requires API keys for AI models (OpenAI, Anthropic, etc.)
• All secrets stored in Docker Desktop secrets
• Never hardcode credentials
• Running as non-root user
• Sensitive data never logged
• Input validation on all parameters
• Timeout protection on external calls

Fabric AI Resources

• GitHub: https://github.com/danielmiessler/Fabric
• Patterns: https://github.com/danielmiessler/Fabric/tree/main/patterns
• Documentation: https://github.com/danielmiessler/fabric/blob/main/README.md

License

GPL-3.0 License - See LICENSE file for details.