
datacenter/mcp_server_for_cisco_aci
Developer toolsA simple MCP (Model Context Protocol) that interacts with Cisco ACI
Overview
This project provides a simple MCP (Model Context Protocol) server that interacts with a Cisco APIC controller. If you'd like to understand how this works in detail, please check out this blog post - Tested with and in Agent mode with Copilot. - The server runs in , intended for local execution. - Exposes two tools for APIC interaction (see app/main.py for details). - Easily configurable via environment variables. 1. in the .env file. 2. If you want Claude or VS Code to run the Python code directly (no container), install UV 3. with Claude or VS Code. For VS Code, create a .vscode/mcp.json file like this in your workspace: 3. Instruct Claude Desktop or VS Code to use it: - See Claude Desktop Quickstart - See VS Code Copilot MCP Servers 4. if you invoke the MCP server with uv run mcp as above. - Use or You can run the server directly using UV, or build a Docker image and run it as a container. If using Docker, adapt the mcp.json config accordingly.
README
MCP (Model Context Protocol) for Cisco APIC
This project provides a simple MCP (Model Context Protocol) server that interacts with a Cisco APIC controller. If you’d like to understand how this works in detail, please check out this blog post
- Tested with Claude Desktop and Visual Studio Code in Agent mode with Copilot.
- The server runs in STDIO mode, intended for local execution.
Features
- Exposes two tools for APIC interaction (see
app/main.pyfor details). - Easily configurable via environment variables.
Setup
-
Specify APIC credentials in the
.envfile. -
If you want Claude or VS Code to run the Python code directly (no container), install UV
-
Register the MCP server with Claude or VS Code.
For VS Code, create a
.vscode/mcp.jsonfile like this in your workspace:{ "servers": { "ciscoApicServer": { "type": "stdio", "command": "C:\\Users\\cpaggen\\.local\\bin\\uv.EXE", "args": [ "run", "--with", "mcp[cli]", "mcp", "run", "C:\\MCP\\app\\main.py" ] } } } -
Instruct Claude Desktop or VS Code to use it:
-
Install MCP client tools locally if you invoke the MCP server with
uv run mcpas above.- Use
uv add "mcp[cli]"orpip install "mcp[cli]"
- Use
Docker Support
You can run the server directly using UV, or build a Docker image and run it as a container. If using Docker, adapt the mcp.json config accordingly.
Note: Local installation of MCP client tools is recommended for debugging the server code.
Screenshots
Below are some screenshots demonstrating the MCP server in action and its integration with Claude Desktop and VS Code:
| MCP Server Registered | Tool Registered in Claude | Claude Tools List |
|---|---|---|
| MCP Server Output | Sample Question in VS Code | Sample Question in Claude |
|---|---|---|
| How to Use ACI Backup |
|---|
These images illustrate the registration process, available tools, and example interactions.
Install
This server does not publish a one-line install command.
Open the repository installation guide New to MCP? Start with the mcpskill.org guide →