any-agent

any-agent's MCP integration exposes a server that allows external MCP tooling to interact with AnyAgent. This MCP server lets you feed prompts, supply tools, and receive agent results through the standard MCP interface, enabling orchestration with other MCP-enabled services and automation pipelines. The server uses the AnyAgent framework to run agent configurations and execute tool calls, returning traces and results that reflect the agent's reasoning and actions. Use this to plug AnyAgent into larger workflows that rely on Model Context Protocol semantics for tooling, evaluation, and agent coordination.

Prerequisites:

• Python 3.11 or newer
• Network access to install Python packages

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

python -m venv venv
source venv/bin/activate  # on Windows use: venv\Scripts\activate

2. Install the AnyAgent package (which includes the MCP server components):

pip install 'any-agent'

3. Run the MCP server (as configured in your mcp_config). Example:

python -m any_agent.mcp_server

4. If you need to customize environment variables, export them before starting the server, or place them in your environment/launch script as required by your deployment.

export ANY_AGENT_MCP_PORT=9000

Notes and tips:

• Python 3.11+ is required. Ensure your environment uses a compatible interpreter.
• The MCP server exposes port configurable via ANY_AGENT_MCP_PORT; adjust as needed for your deployment and firewall rules.
• If you run behind a reverse proxy or in a container, ensure the MCP port is accessible to MCP clients.
• When debugging, check logs for MCP handshake messages, tool invocation results, and trace outputs to understand agent behavior.
• If using multiple frameworks within AnyAgent, ensure the appropriate models/tools are installed and configured in your application code before starting the MCP server.