MCP_server_for_Cisco_ACI

This MCP server provides a lightweight bridge to a Cisco APIC controller. It runs in STDIO mode so it can be used locally or integrated with editor-assisted environments like Claude Desktop or VS Code. The server exposes two tools for APIC interaction via app/main.py; these tools enable common APIC operations from the MCP layer, and are intended to simplify command-and-query workflows against your APIC controller. To connect, register the MCP server with your editor tooling (as shown in the setup example) and ensure the environment is prepared with your APIC credentials. Once registered, you can invoke the tools through Claude or the editor’s MCP integration, passing the appropriate prompts or questions to perform tasks such as retrieving APIC data or executing predefined interactions. The approach is designed to be lightweight and easily configurable through environment variables, so you can tailor credentials and endpoints without modifying the server code.

In practice, you’ll typically access the two APIC tools via the MCP client UI to issue queries or trigger actions against Cisco ACI, with responses streamed back through STDIO. This setup is compatible with ATP workflows in Claude and with VS Code Copilot MCP servers, enabling AI-assisted operations on your APIC data and configurations.

Prerequisites:

• Python installed on your system (recommended Python 3.8+)
• Access to the Cisco APIC controller and credentials
• uv (Astral UV) installed if you plan to run via editor integrations, or ensure you can run the MCP server under your preferred runtime
• Internet access for dependencies and MCP client tools

Installation steps:

1. Clone or download the MCP server repository to your machine.
2. Create and populate an .env file with your APIC credentials and any required configuration values (endpoint, username, password, etc.).
3. Install the MCP client tools (if not already installed):
   • uv add "mcp[cli]" OR pip install "mcp[cli]"
4. Install Python dependencies for the server (if a requirements.txt is provided):
   • python -m pip install -r requirements.txt
5. Register the MCP server with your editor tooling using the snippet provided in the repository (example in the README): create a .vscode/mcp.json file or equivalent in your editor to point to the Python entry point (app/main.py) via uv run mcp, as shown in the README.
6. Run the server using the configured command (example uses uv):
   • uv run --with mcp[cli] mcp run C:\MCP\app\main.py
7. Verify connectivity through the editor or by sending sample prompts to the MCP server and inspecting the responses.

Tips and common considerations:

• Ensure your .env file contains the correct APIC credentials and endpoints before starting the server.
• The server operates in STDIO mode; when integrating with editors, use the provided registration snippet to connect the MCP client to the Python entry point.
• If you opt to run in Docker, adjust the mcp.json configuration accordingly and ensure environment variables are passed into the container.
• The two APIC interaction tools exposed by app/main.py are the primary capabilities; consult the code for exact command names and usage patterns.
• For debugging, install MCP client tools locally to interact with the server directly without the editor, which helps isolate issues related to environment configuration.
• If you migrate from stdio to another runtime, update the mcp_config accordingly to reflect the new command and arguments.