home / mcp / unimus mcp server
MCP for Unimus
Configuration
View docs{
"mcpServers": {
"deployment-team-unimus-mcp": {
"command": "docker",
"args": [
"run",
"-d",
"--name",
"unimus-mcp",
"-e",
"UNIMUS_URL=https://your-unimus.example.com",
"-e",
"UNIMUS_TOKEN=your-api-token",
"-p",
"8080:8080",
"controlaltautomate/unimus-mcp:latest"
],
"env": {
"UNIMUS_URL": "https://your-unimus.example.com",
"UNIMUS_TOKEN": "your-api-token"
}
}
}
}The Unimus MCP Server lets you talk to your Unimus network data through a conversational interface, translating natural language queries into safe, read-only API calls. It enables you to search configurations, inspect backups, map topology, and analyze changes without exposing direct editing capabilities, making complex network insights quick and approachable.
You connect to the MCP server with a client designed to issue queries to the MCP endpoints. Start by ensuring your MCP server is running, then use your client to request data or run analyses such as device searches, backup lookups, or topology mappings. The server exposes read-only access to your Unimus data, so you can ask questions like which devices have had recent configuration changes, locate backups containing specific content, or compare two backups to highlight differences.
Prerequisites you need before installing include Docker (or Python for a direct install path) and a live Unimus instance with a valid API token.
# Docker deployment (recommended)
docker run -d \
--name unimus-mcp \
-e UNIMUS_URL=https://your-unimus.example.com \
-e UNIMUS_TOKEN=your-api-token \
-p 8080:8080 \
controlaltautomate/unimus-mcp:latest# Python installation
git clone https://github.com/Deployment-Team/unimus-mcp.git
cd unimus-mcp
pip install .Environment variables you need to configure the MCP server are: UNIMUS_URL for the full URL to your Unimus instance and UNIMUS_TOKEN for the API token. In Docker, these are provided as -e UNIMUS_URL=... and -e UNIMUS_TOKEN=..., respectively.
- The MCP server is designed for read-only access to facilitate safe conversational queries against your network data. - If you plan to run the server in Docker, ensure you map port 8080 so your MCP client can reach the server. - For production deployments, consider implementing health checks and access controls around the container to protect token visibility.