News for hackers. Stuff that matters.

jimsrand

We built this example as part of a buildathon. It is a real-time voice agent that runs fully offline on iOS (but React Native based so Android will come soon).

VAD, ASR, a small language model, and TTS all run locally on iOS, with a cross-platform architecture that can swap the local model for a frontier cloud LLM when needed.

For now, it's just a button that lets you switch between the on-device SLM and the cloud-based LLM. But we're planning to implement different routing options (e.g. automatic fallback when internet is available, or confidence based, etc.).

Would love to see this remixed. It's designed to be easy to swap components (e.g. use a different VAD, STT, TTS, different language models, etc.).

The implementation is open source. You'll see there's a dependency on the Switchboard SDK. Currently, the free tier is only to 10K MAUs, but we're planning to remove that limit and make most of the core offering a totally free service -- so get in touch if the current model is a blocker for some reason.

Re:Fully local voice AI on iPhone 15 on Thursday March 26, 2026 @08:32AM (#47529662)
good project, going in the right direction... models are getting more efficient, meaning more and more of them can run locally on less powerful phones... we're building many examples in this area, some fully offline, some hybrid
This is a cool use case. Is any of the voice AI running on-device already? If you're planning to do some of the processing on-device and/or cross platform, we're working on some dev tools you might find useful (free to 10K MAUs / cheap after that + lots of open source coming soon). Check out switchboard.audio or feel free to hit me up; always interested in chatting about new use cases in real time voice AI, especially in the consumer space where local/hybrid audio pipelines are useful.
Re:Show HN: Conkoa AI – Voice-First Slack for Construction on Wednesday March 11, 2026 @04:57PM (#47341680)
Definitely still a need for this product. We built an "audio slack" app with some similarities to this a decade ago and it was too early at that time for various reasons. But there's definitely a need for it and now the AI capabilities are far better than what we were working with back then. Good timing.

At some point I suspect you'll want offline functionality / on-device for some aspects of it. Would be interested to chat if you do. Look us up at switchboard.audio -- we are now building dev tools for applications like yours.

Re:Sorca – Voice-first AI therapy companion on Wednesday March 11, 2026 @03:45PM (#47340294)
Cool use case, I have a friend who's starting to experiment in this space as well. Planning to distribute via therapists? Have you explored on-device processing at all (e.g. for latency / cost / privacy reasons?).