Developer builds fully local voice assistant with on-device STT and LLM to avoid voice fingerprinting
wavefnx · x · 2026-11-02
A developer demoed a fully local voice assistant: STT and the LLM both run on-device, with internet access, web search, image display, and tool calling hooked in. Responses play over a speaker (not yet wired up) with on-screen transcripts.
The pitch: keeping voice and conversation data local means no more voice fingerprinting. A live test of "Bot, book my helicopter for 5:00pm" showed the mic and STT pipeline working cleanly.
Related event: Developer Builds DIY Desktop AI Robot with Local LLM and Voice(3 posts)→
More from coding & agent
- Phone Harness MCP lets any MCP agent control real cloud iPhones and Androids — Scobleizer · 2026-11-03
- LangChain on late-2026 context engineering: browser use, long-term goals, async work — hwchase17 · 2026-11-03
- MotherDuck launches Flight Explainer: live charts for agent-built Python pipelines — josh_wills · 2026-11-03
- Devin alternatives for enterprise teams: Factory, Copilot, Codex compared — Available_Bit4986 · 2026-11-03
- Vibe coding's free ride is ending: one simple app racked up a $13,594 bill — michalmalewicz · 2026-11-03
- Using AI agents for Java legacy migrations at banks and insurers — Fit_Satisfaction541 · 2026-11-03