Voicetypr 2.0 - Stop typing, Start talking to your apps and agents

Hey Product Hunt 👋 I'm Moin, the solo founder of Voicetypr. I first launched here a while back — today I'm back with 2.0, the biggest rebuild yet. Why I built it I type 12+ hours a day, and paying a monthly subscription for basic dictation never sat right with me. Most voice tools were cloud-only (your voice leaves your machine), Mac-only, or locked behind a recurring fee. I wanted one that was fast, private, ran on my own machine, and that I'd pay for once. So I built it. What changed for 2.0 While building v1, the AI-agent wave happened — people now run personal agents (OpenClaw, Hermes, and friends) on their own machines. Those agents could read and act, but they couldn't *hear* you without piping your voice to a paid cloud API. So 2.0 became a ground-up rebuild (every subsystem) around one idea: give your agents ears, locally. • A new Agent CLI + local API — any agent or script can pull your voice as text/JSON, on-device. No per-minute voice API, no audio sent to the cloud. • Remote sharing — let a GPU desktop transcribe for a lighter laptop over your own Wi-Fi. • Way more powerful formatting — 6 modes (clean dictation, writing, notes, message, code…) plus per-app rules. • Optional cloud engines if you want them (Soniox, OpenAI, Groq, Deepgram, Cohere) — local stays the default. Plus a new native hotkey engine, rich searchable history, and a lot of recording-reliability work. What didn't change: transcription is still local and offline by default, and it's still a one-time purchase. No subscription. Try it free for 3 days, no card — on Mac and Windows. I'm here all day and would genuinely love your feedback and questions. 🙏 — Moin