Hey Product Hunt! 👋 I built SyncDialect because most real-time voice translators depend on the cloud. That creates three problems I wanted to tackle: connectivity, latency and privacy. SyncDialect takes a different approach: the core translation pipeline runs directly on your Android device. It combines Google's Gemma + LiteRT for on-device AI, WebRTC voice activity detection, and streaming text-to-speech to create a natural voice-to-voice translation experience. Once the required model is downloaded, the core translation workflow doesn't need an internet connection and conversations don't need to be sent to a translation server. The goal is simple: you should be able to talk to someone in another language whether you're travelling without mobile data, somewhere with unreliable connectivity, or simply don't want a private conversation processed in the cloud. I've been especially focused on reducing the delay between someone finishing a sentence and hearing the translation. The current implementation is designed around a low-latency pipeline rather than treating voice translation as a series of separate manual steps. SyncDialect is still evolving, and I'd really value feedback from the Product Hunt community. I'd particularly love to hear: • Which language pairs would you use most? • Where would offline voice translation be useful to you? • What would make you choose an offline translator over Google Translate or other cloud-based alternatives? Thanks for checking out SyncDialect! 🚀

SyncDialect

Real-time voice translation that never leaves your phone

Real-time voice translation that never leaves your phone