Our product · Deep focus
Quell: a 40 Hz binaural focus app, generated in the browser, no backend
A focus app that generates a 40 Hz binaural beat, a neural-voice coach, and AI music on the device. No accounts, no server, works offline.
Binaural beat
130 / 170 HzLofi Night
Track 2 of 4
The story behind
It started as one screen. Tap Begin, focus, End. The idea was easy. The audio was not. A binaural beat is two pure tones a fixed distance apart, and the distance has to be exact: 130 Hz in one ear and 170 in the other gives the 40 Hz beat the brain is offered. The carrier matters too. The textbook choice is around 400 Hz, but it sounds thin and piercing, so Quell uses a deep, warm 150 Hz instead. The browser hands you raw oscillators, not a play-a-40-Hz-beat button, so the whole thing is two sine waves and a channel merger, built by hand.
Then the unglamorous parts showed up. Browsers block audio that starts without a tap, so the beat, the coach, and the music all have to begin inside the same Begin gesture. Three independent layers have to pause as one. And because the brief was that it should work on a plane, the app ships with no backend: the coach lines are pre-rendered neural speech, bundled and precached, and the music is cached on first play. Honest framing was a requirement, so the app says out loud that the evidence for binaural beats is mixed. That is the whole job, the quiet correctness, not the demo.
Business value
- No backend means nothing to breach and nothing to run.
- The beat is synthesized live, so the app is small and loads instantly.
- The coach voice is pre-rendered once, not generated per session, so there is no per-use cost and no network round-trip.
- Everything stays on the device. No account, no tracking, no analytics.
Project scope
- A Web Audio engine for the 40 Hz binaural beat, with an isochronic fallback for speakers.
- A count-up session timer with no fixed length; ending the session is the complete event.
- A three-layer soundscape mixer (beat, coach, optional music), each with its own volume.
- A pre-rendered neural-TTS coach and an AI-generated music library, both owned, not streamed.
- An offline-first PWA, plus iOS and macOS builds through Capacitor.
Deliverables
- The live app at runquell.com, installable as a web app.
- Quell on the App Store for iOS and macOS.
- The binaural and isochronic audio engine, built on the Web Audio API.
- A coach-voice pipeline: fixed lines pre-rendered with OpenAI neural TTS, bundled and precached.
- A 16-track music library generated with Google Lyria, cached on first play.
Tech stack
Frequently asked
Can I try Quell?
Yes. It is live at runquell.com and installs as a web app, and it is on the App Store for iOS and macOS under Quell Deep Focus. The beat and the count-up timer are free. The coach voice and the music are the paid part.
Do binaural beats actually work?
Honestly, the evidence is mixed. Meta-analyses put the effect somewhere around a small-to-moderate g of 0.4, not the dramatic claims some apps make, and Quell says so in plain text inside the app. We built it because a quiet, honest focus aid is useful even if the beat is only part of why. It is also not the MIT 40 Hz gamma-flicker research, which is a different thing.
Can you build something like this for us?
Yes. The pieces transfer: real-time audio in the browser, an offline-first PWA with a precache strategy, a pre-rendered neural-voice pipeline, and a native wrapper through Capacitor. If you have an app that should work without a backend or without a network, this is the shape of it.
Why no accounts or backend?
Because the product is a ritual, not a dashboard. Nothing it does needs a server: the audio is generated on the device, settings live in local storage, and the coach and music are bundled or cached. Less to breach, less to run, and it works with the Wi-Fi off.
Have a workflow that needs this?
Tell us the shape of the problem. Scoped estimate, usually within 3 to 5 business days. No card, no obligation.
Estimate this buildor email business@highcraft.ioMore work
AI Lab Analysis
AI-assisted intake reads raw lab PDFs and drafts plain-language summaries a provider reviews before the visit.
EMR & Chart Notes
An EMR and patient portal built in .NET and Azure, developed under real regulatory scrutiny on a normal sprint cadence.
Clinical AI Insights
Insights compiled from a patient's recent notes, latest lab, and intake, with the sources it used and the ones it could not.