Strands MCP Server

A Model Context Protocol (MCP) server that provides semantic search capabilities for the Strands Agent SDK documentation. This server enables AI assistants and development tools to access up-to-date Strands SDK documentation through intelligent search and retrieval.

Features

• Semantic Documentation Search: Fast, intelligent search across Strands SDK documentation
• Bundled Cache: Pre-built documentation cache for immediate availability
• Background Updates: Automatic documentation updates with configurable intervals
• Fast Startup: Instant search capabilities using bundled cache
• GitHub Integration: Fetches latest documentation from Strands repositories
• Health Monitoring: Built-in health checks and error handling
• MCP Compatible: Works with any MCP-compatible client (Claude Desktop, Kiro IDE, etc.)

Quick Start

Installation

# Install from PyPI
pip install strands-mcp-server

# Or install with uv (recommended)
uv add strands-mcp-server

Basic Usage

With Claude Desktop

Add to your Claude Desktop MCP configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "strands-docs": {
      "command": "python",
      "args": ["-m", "strands_mcp.main"],
      "env": {
        "GITHUB_TOKEN": "your_github_token_here"
      }
    }
  }
}

With Kiro IDE

Add to your Kiro MCP configuration (.kiro/settings/mcp.json):

{
  "mcpServers": {
    "strands-docs": {
      "command": "uvx",
      "args": ["strands-mcp-server@latest"],
      "env": {
        "GITHUB_TOKEN": "your_github_token_here"
      },
      "disabled": false,
      "autoApprove": ["search_documentation", "list_documentation"]
    }
  }
}

Direct Usage

# Run the MCP server directly
python -m strands_mcp.main

# Or with uvx
uvx strands-mcp-server@latest

Configuration

Environment Variables

GitHub Token Setup

A GitHub token is recommended to avoid API rate limits when fetching documentation.

Creating a GitHub Token

1. Go to GitHub Settings > Developer settings > Personal access tokens
2. Click "Generate new token (classic)"
3. Give it a descriptive name like "Strands MCP Server"
4. Select scopes: public_repo (for accessing public repositories)
5. Click "Generate token"
6. Copy the token (you won't see it again!)

Setting the Token

Quick Setup (Recommended):

# Use the built-in setup utility
strands-setup-token

# Or with pip/uvx
pip install strands-mcp-server
strands-setup-token

Manual Setup:

# Option 1: Environment variable
export GITHUB_TOKEN=ghp_your_token_here

# Option 2: In your shell profile (.bashrc, .zshrc, etc.)
echo 'export GITHUB_TOKEN=ghp_your_token_here' >> ~/.zshrc

# Option 3: In MCP configuration (see examples above)

For Production/CI:

• Use GitHub repository secrets (see GitHub Secrets Setup Guide)
• Use environment-specific secret management
• Never commit tokens to source control

Available Tools

The MCP server provides the following tools:

search_documentation

Search Strands SDK documentation using semantic search.

Parameters:

• query (string): Search query
• limit (integer, optional): Maximum results (default: 10, max: 50)
• min_score (number, optional): Minimum relevance score (default: 0.3)

Example:

{
  "name": "search_documentation",
  "arguments": {
    "query": "how to create a multi-agent workflow",
    "limit": 5,
    "min_score": 0.5
  }
}

list_documentation

Browse available documentation sections and documents.

Parameters:

• section_filter (string, optional): Filter by section name
• limit (integer, optional): Maximum results (default: 20, max: 100)

Example:

{
  "name": "list_documentation",
  "arguments": {
    "section_filter": "multi-agent",
    "limit": 10
  }
}

Health Check Tools

• health_check: Check server health status
• readiness_check: Check if server is ready to serve requests
• liveness_check: Check if server is alive and responsive

Development

Prerequisites

• Python 3.11 or higher
• uv (recommended) or pip
• Git

Setup

# Clone the repository
git clone https://github.com/strands-agents/strands-mcp-server.git
cd strands-mcp-server

# Install dependencies with uv
uv sync

# Or with pip
pip install -e ".[dev]"

# Set up pre-commit hooks (optional)
pre-commit install

Setting Up GitHub Token

For development, you'll need a GitHub token to avoid rate limits:

# Use the setup utility (recommended)
uv run strands-setup-token

# Or set manually
export GITHUB_TOKEN=your_token_here

Building Documentation Cache

The server includes a bundled documentation cache, but you can rebuild it:

# Build with GitHub token (recommended)
GITHUB_TOKEN=your_token python scripts/build_bundled_cache.py --force

# Build without token (rate limited)
python scripts/build_bundled_cache.py --force

# Verbose output
python scripts/build_bundled_cache.py --force --verbose

Running Tests

# Run all tests
uv run pytest

# Run with coverage
uv run pytest --cov=strands_mcp --cov-report=html

# Run specific test categories
uv run pytest tests/unit/          # Unit tests only
uv run pytest tests/integration/   # Integration tests only

Code Quality

# Format code
uv run black src/ tests/

# Lint code
uv run ruff check src/ tests/

# Type checking
uv run mypy src/

CI/CD and GitHub Token Access

Repository Secrets

For automated builds and deployments, configure these GitHub repository secrets:

1. Go to your repository Settings > Secrets and variables > Actions
2. Add the following secrets:

GitHub Actions Workflows

The repository includes automated workflows:

Build and Test (ci.yml)

• Runs on every push and pull request
• Tests across multiple Python versions
• Builds documentation cache
• Runs code quality checks

Publish to PyPI (publish.yml)

• Runs on new releases
• Builds and publishes package to PyPI
• Uses trusted publishing (no manual tokens needed)

Update Documentation Cache (update-cache.yml)

• Runs daily to update bundled documentation
• Creates pull requests with updated cache
• Ensures documentation stays current

Manual Workflow Triggers

You can manually trigger workflows:

# Trigger cache update
gh workflow run update-cache.yml

# Trigger build and test
gh workflow run ci.yml

# Trigger publish (requires release)
gh workflow run publish.yml

Architecture

Cache Hierarchy

The server uses a multi-level cache system:

1. Bundled Cache: Pre-built cache included in the package
2. User Cache: Local cache updated in the background
3. Memory Cache: In-memory cache for frequently accessed data

Documentation Sources

• Primary: Strands Agent SDK Documentation
• Fallback: Bundled cache included with the package
• Updates: Automatic background updates from GitHub

Troubleshooting

Common Issues

"No documentation found" or empty search results

• Check if GitHub token is configured correctly
• Verify internet connectivity
• Try rebuilding the cache: python scripts/build_bundled_cache.py --force

Rate limit errors

• Add a GitHub token to your configuration
• Check token permissions (needs public_repo scope)
• Wait for rate limit to reset (usually 1 hour)

MCP connection issues

• Verify MCP client configuration
• Check server logs for error messages
• Ensure Python path is correct in MCP config

Import or module errors

• Reinstall the package: pip install --force-reinstall strands-mcp-server
• Check Python version compatibility (3.11+ required)
• Verify virtual environment activation

Debug Mode

Enable debug logging for troubleshooting:

# Environment variable
export LOG_LEVEL=DEBUG

# Or in MCP configuration
{
  "env": {
    "LOG_LEVEL": "DEBUG"
  }
}

Getting Help

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Documentation: Strands Agent SDK Docs

Contributing

We welcome contributions! Please see our Contributing Guide for details.

Development Workflow

1. Fork the repository
2. Create a feature branch: git checkout -b feature/amazing-feature
3. Make your changes
4. Add tests for new functionality
5. Run tests and ensure they pass
6. Commit your changes: git commit -m 'Add amazing feature'
7. Push to your branch: git push origin feature/amazing-feature
8. Open a Pull Request

Code Standards

• Follow PEP 8 style guidelines
• Add type hints for all functions
• Write docstrings for public APIs
• Include tests for new features
• Update documentation as needed

License

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

Changelog

See CHANGELOG.md for a detailed history of changes.

Built with ❤️ for the Strands Agent SDK community