mkinf

The mkinf MCP server is a Python-based SDK server that exposes the functionality of the mkinf package as an MCP-compatible service. It enables you to discover, configure, and run AI agents from the mkinf hub within your MCP-powered workflows. Through the server you can pull agents, pass environment configuration, and leverage the SDK to interact with AI capabilities in a uniform, pluggable manner.

To use it, first obtain an API key from the mkinf hub and start the MCP server as configured. Within your MCP pipelines, you can invoke agents by their identifiers (for example, pulling an agent like ScrapeGraphAI/scrapegraphai) and provisioning any required environment variables. The included example demonstrates how to pull an agent with a specified environment mapping, so you can tailor model selections and credentials (such as API keys) per agent usage. Remember that agents may require additional environment variables described in their documentation, so make sure those are provided to the MCP server when invoking the agent.

Prerequisites:

• Python 3.8+ installed on your system
• Access to install Python packages (pip)
• Optional: virtual environment tool (venv, virtualenv)

Step-by-step:

1. Create and activate a virtual environment (optional but recommended):

python -m venv venv
# Windows
venv\Scripts\activate
# macOS/Linux
source venv/bin/activate

2. Install the mkinf SDK from PyPI:

pip install mkinf

3. Verify installation (example):

python -m mkinf --help

4. Configure your MCP server (see mcp_config). Save your API key in a secure way, e.g., environment variable MKINF_API_KEY.

5. Start the MCP-compatible server via your MCP orchestration (as described in your MCP deployment). The example configuration uses a Python module approach (python -m mkinf).

Notes and tips:

• API credentials: Store MKINF_API_KEY securely, and do not hard-code it in code or config files. Use environment variables or a secrets manager.
• Agent configuration: When pulling an agent, you may need to provide environment mappings (env) for model identifiers, keys, or other per-agent settings. Align these with the agent's own documentation.
• Compatibility: The mkinf MCP server is Python-based. If you deploy in a containerized environment, ensure Python dependencies are installed and network access to the mkinf hub is available.
• Debugging: If you encounter authentication or rate-limit issues, check that the API key has the necessary permissions and that you are using the correct account/organization context in the hub.
• Updates: Regularly update the mkinf package to receive the latest agents and capabilities.