Launching today on PH: I built a voice synthesis app that runs entirely on your Mac (no cloud, no API keys)
Today I’m launching Sovereign Voice on Product Hunt.
I built it because I was paying a cloud voice service per character to synthesize narration for short videos. One day it hit me: my MacBook Pro has a dedicated neural accelerator sitting idle 90% of the time. Why am I sending my scripts to a server, paying for each character, and trusting a third party with my voice data?
So I built the opposite architecture:
- The TTS model (Qwen3-TTS 0.6B, Apache-2.0) runs via MLX on the Metal GPU
- Model weights live on your disk (~1.9GB, one-time download)
- After that: fully offline. Pull the ethernet cable, it still works.
- No account, no API key, no per-character billing
Free tier: 25 generations, 1 voice.
Pro: $19 one-time (unlimited everything, voice cloning, studio mode).
I chose one-time pricing over subscription because the compute happens on the user’s hardware — there’s no ongoing server cost to justify a recurring charge.
It’s Mac-only (Apple Silicon) for now. The M4 does synthesis in 5-9 seconds.
Would love feedback from anyone here who works with audio content, voiceover, or cares about local-first software. Happy to answer questions about the architecture, the model choice, or the pricing.
Update: PH launch is live now! Check it out at sovereign-voice-landing.pages.dev — the Product Hunt page should be live. Would love your support and feedback.