OTOBO AI MCP server

This is the OTOBO MCP server. It exposes parts of the OTOBO webservices to an LLM as part of the otobo-ai-services. It assumes a running OTOBO instance with otobo-ai package installed.

Build

To build the server

docker compose build

To build the .venv (only needed if you want to run the integration test client client.py)

# install python uv (here:ubuntu), or use pip
sudo snap install astral-uv --classic

uv venv
source .venv/bin/activate
uv sync

Run MCP Server

docker compose up -d

Integration Test

The integration test is implemented in the client.py.

The MCP uses a OTOBO SessionID bound to the MCP server IP address. Obtain a session ID:

docker exec ti mcp curl -vk \
  http://web:5000/otobo/nph-genericinterface.pl/Webservice/Session/Session \
  --header "Content-Type: application/json" \
  --data '{"UserLogin" : "root@localhost", "Password" : "root"}'

Run the client with this session id:

uv run client.py <SessionID>