Case study: the computer that introduces itself

Most software greets you with a config file. Ours says hello: during installation the machine speaks, introduces itself and walks you through setup like a colleague — multilingual, with automatic language detection.

The whole voice loop (microphone → voice-activity detection → local Whisper transcription → LLM → speech) runs on your own hardware, fully offline. We verified it end-to-end even on an aging NVIDIA Quadro M5000 workstation card — no cloud, no audio ever leaving the device.

Adam and Eva are just the default pair — the first man and the first woman, showing the assistant can be either gender. Name your assistant anything, pick any voice, and get as many voices as your plan has agents. For accessibility users, voice is a first-class input equal to the keyboard — not an afterthought.