nic01asFr/docs-mcp-server

Docs MCP Server

The Docs MCP Server provides seamless integration between Claude and Docs instances through the Model Context Protocol (MCP). It enables Claude to interact with collaborative documents, manage access permissions, and leverage AI-powered features directly within the Docs ecosystem.

✨ Key Features

📝 Complete Document Management

• 📄 Create, read, update, and delete documents
• 🌳 Navigate hierarchical document structures
• ↔️ Move and reorganize documents in trees
• 📋 Duplicate documents with or without permissions
• ⭐ Manage favorites and restore from trashbin

✏️ Document Content Editing (NEW in v0.2.0)

• 📖 Read document content as plain text
• ✍️ Update documents with text or markdown
• 🔄 Apply AI transformations directly to documents
• 🌍 Translate document content automatically
• 🔧 Yjs (CRDT) format support for collaborative editing

👥 Advanced Access Control

• 🔐 Grant and revoke user permissions (reader, editor, administrator, owner)
• 📧 Send email invitations to external users
• 🔍 Search for users across the platform
• 📮 Manage pending invitations

🤖 AI-Powered Features

• ✍️ Text correction and grammar checking
• 🔄 Content rephrasing and summarization
• 🌍 Multi-language translation support
• ⚡ Custom AI transformations

📚 Version History

• 📖 Browse document version history
• 🔍 Retrieve specific version content
• 📊 Track changes over time

🔌 MCP Integration

• 31 Tools: Comprehensive set of operations including content editing
• 4 Resources: Real-time data access
• Type Safety: Full TypeScript-style type hints
• Error Handling: Robust error management
• Yjs Support: Native collaborative document format

🚀 Installation

From PyPI (Recommended)

pip install docs-mcp-server

From Source (Development)

git clone https://github.com/nic01asFr/docs-mcp-server.git
cd docs-mcp-server
pip install -e \".[dev]\"

⚙️ Configuration

Environment Variables

export DOCS_BASE_URL=\"https://your-docs-instance.com\"
export DOCS_API_TOKEN=\"your-api-token\"
export DOCS_TIMEOUT=30          # Optional: request timeout in seconds  
export DOCS_MAX_RETRIES=3       # Optional: maximum retry attempts

Verify Configuration

docs-mcp-server --config-check

Output:

✓ Configuration loaded successfully
  Base URL: https://docs.example.com
  Token: ****-token-****-1234
  Timeout: 30s
  Max retries: 3
✓ API connection successful
  Authenticated as: user@example.com
  User ID: user-123

🛠️ Usage

Command Line

# Start with environment variables
docs-mcp-server

# Start with explicit configuration  
docs-mcp-server --base-url https://docs.example.com --token your-token

# Start with custom server name
docs-mcp-server --name my-docs-server

# Start with verbose logging
docs-mcp-server --verbose

Python Module

python -m docs_mcp_server

Programmatic Usage

import asyncio
from docs_mcp_server import DocsServer

async def main():
    server = DocsServer(
        base_url=\"https://docs.example.com\",
        token=\"your-token\",
        server_name=\"my-docs-server\"
    )
    await server.run()

if __name__ == \"__main__\":
    asyncio.run(main())

📋 Available Tools

📊 Resources

💡 Examples

Basic Document Operations

from docs_mcp_server import create_client

async def example():
    async with create_client() as client:
        # Create a document
        doc = await client.create_document(
            title=\"Project Proposal\",
            content=\"# Project Overview\\n\\nThis is our new project...\"
        )
        
        # Grant access to a colleague
        await client.grant_access(
            document_id=doc.id,
            user_email=\"colleague@example.com\",
            role=\"editor\"
        )
        
        # Use AI to improve content
        improved = await client.ai_transform(
            document_id=doc.id,
            text=\"This text needs improvement\",
            action=\"rephrase\"
        )
        print(f\"Improved text: {improved.result}\")

MCP Server Integration

import asyncio
from docs_mcp_server import DocsServer

async def main():
    server = DocsServer(
        base_url=\"https://docs.example.com\",
        token=\"your-token\",
        server_name=\"company-docs\"
    )
    await server.run()

asyncio.run(main())

Error Handling

from docs_mcp_server import DocsAPIClient, DocsError, DocsNotFoundError

async def robust_example():
    try:
        async with DocsAPIClient() as client:
            doc = await client.get_document(\"non-existent-id\")
    except DocsNotFoundError:
        print(\"Document not found\")
    except DocsError as e:
        print(f\"API error: {e.message}\")
    except Exception as e:
        print(f\"Unexpected error: {e}\")

🧪 Development

Setup Development Environment

git clone https://github.com/nic01asFr/docs-mcp-server.git
cd docs-mcp-server
pip install -e \".[dev]\"
pre-commit install

Run Tests

# Run all tests
pytest

# Run with coverage
pytest --cov=docs_mcp_server --cov-report=html

# Run specific test file
pytest tests/test_client.py -v

Code Quality

# Linting and formatting
ruff check src/ tests/
ruff format src/ tests/

# Type checking
mypy src/docs_mcp_server

# Security scanning
bandit -r src/
safety check

Documentation

# Serve documentation locally
mkdocs serve

# Build documentation
mkdocs build

🏭 Production Ready

✅ Quality Assurance

• 🧪 Comprehensive test suite with >95% coverage
• 🔒 Type safety with mypy and pydantic
• 🧹 Code quality with ruff and pre-commit hooks
• 🛡️ Security scanning with bandit and safety
• 📊 Performance monitoring and optimization

🔐 Security

• 🔑 Secure API token management
• 🌐 HTTPS-only communication
• ✅ Input validation and sanitization
• 💾 No sensitive data storage
• 📋 Comprehensive security documentation

🚀 CI/CD

• 🔄 Automated testing on multiple Python versions (3.10-3.12)
• 📦 Automated PyPI publishing on releases
• 🔍 Security vulnerability scanning
• 📈 Performance regression testing

📚 Documentation

• 📖 Comprehensive API documentation
• 💡 Usage examples and tutorials
• 🤝 Contribution guidelines
• 🔒 Security policy
• 📋 Detailed changelog

🔗 Links

• 📦 PyPI Package
• 📖 Documentation
• 🐙 GitHub Repository
• 🐛 Issue Tracker
• 💬 Discussions

📄 License

This project is licensed under the MIT License - see the file for details.

🤝 Contributing

We welcome contributions! Please see our for details.

Contributors

<a href="https://github.com/nic01asFr/docs-mcp-server/graphs/contributors\"> <img src="https://contrib.rocks/image?repo=nic01asFr/docs-mcp-server\" />

🙏 Acknowledgments

• Model Context Protocol for the MCP specification
• La Suite Numérique and the DINUM team for creating the Docs platform
• All contributors and users of this project