magic

Magic is an open-source, all-in-one AI productivity platform that combines multiple AI-assisted capabilities into a single ecosystem. The server exposes tools for orchestrating AI workflows (Magic Flow), managing intelligent conversations (Magic IM), and integrating with enterprise communication channels and data sources. With the Magic MCP server running, you can start building, deploying, and running AI agents that can plan, execute tasks, and autonomously correct errors across a matrix of components. The platform emphasizes interoperability among components, enabling agents to trigger workflow steps, query knowledge bases, and interact with users or other services through a unified API surface.

Prerequisites:

• Python 3.8+ and pip
• Git
• Optional: PostgreSQL/Redis if you plan to use persistent storage and caching

Step-by-step:

1. Clone the repository: git clone https://github.com/dtyq/magic.git cd magic

2. Create a virtual environment and install dependencies: python -m venv venv source venv/bin/activate # on Unix or .\venv\Scripts\activate # on Windows pip install -r requirements.txt

3. Configure environment variables (create a .env file or export variables): export OPENAI_API_KEY=your-openai-api-key export DATABASE_URL=postgres://user:pass@host:port/dbname export REDIS_URL=redis://localhost:6379

4. Run the MCP server: python -m magic.server

5. Optional: run with a specific config file or docker if provided by the project docs. Refer to the repository's documentation for any project-specific command-line options.

Environment variables and configuration details may vary by deployment. Common variables include OPENAI_API_KEY for model access, DATABASE_URL for persistent storage, and REDIS_URL for caching/queueing. If you encounter connection or authentication issues, verify network access to your datastore and that API keys are correctly set. When running in production, consider configuring TLS, proper secret management, and rate-limiting on API endpoints. If the server exposes admin or diagnostic endpoints, secure them behind authentication and restrict access via firewall rules.