TL; DR
🌐 https://ariatype.com
💻 https://github.com/SparklingSynapse/AriaType
👋 Hey Product Hunt!
Original story:
I am a CS developer with 10 years of experience. This spring festival, I was working on a web coding project.
About 3 days in, I realized something: the biggest challenge wasn’t the cost of AI subscriptions. It was my stamina.
The constant cycle of writing prompts, correcting AI outputs, and changing contexts was exhausting. So I made myself a voice input tool.
Why make it yourself?
I looked at Typeless. The subscription was steep – almost identical to my AI plan. I wasn’t willing to pay that much for what I could make for myself. (But it’s great, obviously)
Two hours. MVP.
LLaMA CPP + whisper. Rough around the edges, but functional. Good enough for personal use.
Then the real work began:
After sharing with colleagues after the holiday, I got real feedback. Some of it stung, but it was all worth it. So I spent the next weekend polishing it.
What AriaType 0.1 does:
– Spatial STT models – Whisper based. Vesper for English, Sense Voice for CJK (Alibaba)
– AI Polish – Local miniature models for grammar correction, filler removal
– Cloud Mode – Bring your own AI subscription. No separate payment.
– Noise reduction and silence detection – omits silence to save costs.
– 100+ languages.
– Privacy by default — Voice data never leaves your machine.
I am proud of:
– 100% open source
– Everything runs locally by default.
– Works in any app via global hotkey.
– Minimal marks on Apple silicone
Honesty Challenge:
The hardest part wasn’t MVP. It was after 80% – converting the AI reliably and expanding the growing code base. This is called “harness engineering”.