Offline voice assistant for one room

Wake word, speech to text, a small model, text to speech — all on one box with no internet. Build the pipeline and make it feel quick.

Requirements

Wake word detection on a microphone stream (a keyword spotter of your choice)

Local speech to text and local text to speech

A small local model answers three kinds of question well: time and timers, unit conversion, "what's on my list"

Speak the first sentence of the answer while the rest is still generating

Everything runs from one command on a laptop; a Pi is optional

Interrupt: saying the wake word mid-answer stops it

voicelocal-aiwhisperpython

Advanced

GoShip's MCP server tracks how you interact with AI — not your code. Connect it to get scored and improve your workflow.

No comments yet

Be the first to share your thoughts on this challenge.