One assistant for your homelab

Ambrogio runs on your server, next to the services you already host. Each agent reaches one of them through the Model Context Protocol, and the orchestrator picks the right agent for each request.

The services of ambrogio-homelab-example. Logos are trademarks of their owners.

What it does

• Voice in every room

  A small device with a microphone and a speaker waits for its wake word and answers out loud. Speech is recognised and spoken on your server, and audio is never stored.

  Voice →

• Chat from the browser or the terminal

  Type to Ambrogio from the web chat or a terminal. Every conversation is shared: a question asked in the kitchen can go on in the browser.

  Conversations →

• Multi-agent by design

  Each agent you configure handles one job, such as the house, music, weather or your documents, with its own instructions and its own model. An orchestrator picks the right one for each request.

  Agents →

• The services you already host

  Agents use your services through the Model Context Protocol: Home Assistant, Music Assistant, OpenCloud, or any other MCP server.

  Connect an MCP server →

• Your choice of LLM

  Each agent uses an LLM from a cloud provider or one that runs on your network. With every model at home, nothing leaves it.

  Choose the LLM provider →

• Runs on your server

  The backend runs with Docker on a home server and keeps every conversation in its own database, ready to search later.

  Before you install →