Every snore app I tried flagged the humidifier, the fan, and the dog. So I built one that doesn't. Snore IQ runs 8 hours of AI inference entirely on your phone. No cloud, no audio uploaded, ever — and 8-13% battery overnight. It cuts real snoring out of ambient noise, tracks trends and positional patterns, and exports reports for your doctor. 3 free nights, then one-time pricing. No subscription: local AI means near-zero cost per user.
Tines The single, secure environment for agents, apps, and automations.
Promoted
Maker
📌
Hey Product Hunt 👋
I built Snore IQ because my wife told me I snore and I didn't believe her. The recordings disagreed.
Every app I tried failed the same way: it flagged the humidifier, then the fan, then the dog, then my wife breathing. I'd wake up to a "snoring score" made almost entirely of appliance noise. The ones that worked better did it by uploading eight hours of bedroom audio to a server, which isn't a trade I wanted to make.
So everything in Snore IQ runs on your phone. No cloud, no server, no audio leaving the device. There's no backend to breach because there's no backend.
The question I always get first: what about the battery? Fair — 8 hours of continuous inference sounds brutal. My first working version would have killed a phone overnight. Months of work on duty cycling, buffer sizing, and wake-lock behavior got it to 8–13% overnight depending on the device. That was most of the engineering.
The part I didn't expect: a few months of my own data sent me to a sleep study. Diagnosis was airway collapse. I sleep with a mandibular device now. The app changed my health before it ever shipped — which is why it exports reports a clinician can actually read, and why STOP-BANG screening will tell you to see a doctor instead of suggesting a pillow when your risk profile warrants it.
Pricing is one-time, not a subscription. Local inference means my cost per user is basically zero, so there's no monthly bill to pass on to you. Three free nights to try it, then a 90-day pass or lifetime access — early lifetime is capped at the first 1,000 users.
Happy to go deep on the on-device inference, the noise rejection work, or the battery optimization. I'll be here all day. 🙏