Speak to an AI that hears you, thinks, and answers out loud — entirely on your own GPU. It can act on your machine too: read your files, run local tools, take actions. No call leaves the box. No per-minute meter.
Cloud voice platforms can call functions — but only as webhooks back to a server on the public internet. Their brain lives in the cloud, so they can’t read your files, open your apps, or reach a service on localhost. Bare Metal Voice runs the whole loop — speech, model, and tool execution — on the same box. So a “tool call” is a local function on your machine: instant, private, and able to touch things no cloud agent can reach.
Whisper runs locally and transcribes as you talk. Your voice is never streamed to a server.
Any model in the catalog, with tool-calling wired below the model so the agent loop actually holds.
Read a file, run a command, hit a localhost service — local execution a cloud webhook can’t reach.
Natural text-to-speech renders the reply on your own GPU. The whole round-trip stays offline.
Your voice, your model, your machine — the call never leaves the box.