On-Device Voice AI
Rush Voice Butler
100% speech privacy · <280ms latency
Demo video unavailable
01 Problem
[problem: why speech had to stay on the device: privacy, cloud dependency, latency]
02 Approach
Always-listening voice butler running fully on-device: Vosk STT + Piper neural TTS, Spotify Connect, and a Telegram & n8n bridge. 100% on-device speech privacy.
03 Result
Delivered 100% private, zero-cloud speech processing directly on edge Raspberry Pi 4 hardware, operating reliably without external API dependencies or data leakage risks.
100% speech privacy · 99.9% local uptime · <280ms local speech pipeline
- 100%On-device speech privacy
- <280msLocal pipeline latency
- 99.9%Local hardware uptime
- 0 packetsZero audio sent to cloud