On-device voice isolation · Android

Every voice.
Crystal clear.

AI that isolates specific voices in real time — right on your Android phone, with zero cloud. Voice Choice gives you the right to choose what you hear.

Get early access Try the Sound Lab
<100ms
Latency
100%
On-device
0
Cloud data
#1
Innovation of the Year Dalarna
The Problem

Noisy café. Loud classroom.

Humming fans. Crucial meetings.

The voice you're trying to listen to — buried.

It's called the cocktail party problem — the brain's inability to isolate a single conversation stream in chaotic environments. For most, this happens automatically. But for the 15–20% with ADHD, autism, or sensory sensitivities, there is no such filter. Every noisy room drains cognitive capacity.

The market's answer? Noise-canceling headphones that block everything. Apple, Sony, Bose craft shields that protect us from the noise — a huge step in the right direction. Voice Choice takes the next step: it makes what matters audible.

15–20%
of the population live with neurodivergence affecting auditory processing
Source: Doyle 2020, British Medical Bulletin
17%
of approved occupational illnesses in Sweden are linked to noise exposure
Source: Luleå University of Technology
5–10 bn
SEK (Swedish kronor) per year in societal costs from road-traffic noise (SIKA estimate)
Source: Chalmers 2013, citing SIKA

How it works

From voice sample to voice isolation

Three steps — and all of it happens on your phone.

Record

Ask the person to speak for ten seconds. The voice becomes a fingerprint — unique to that one human.

Build the profile

The AI turns the sample into a voice profile, stored locally on the device. No cloud, no upload, no fuss.

Listen

The engine captures the audio, filters in real time under 100 ms — and hands back a single voice: the one you chose.

Works as a software layer on top of hearing aids and active noise cancellation.

Who it's for

Built for those who never got to choose.

The classroom

Hear the teacher through the noise — wherever you sit in the room.

Neurodivergent minds

ADHD, autism, and sensory sensitivity: loud spaces become manageable when you choose what gets through.

Open offices

One conversation in focus — the whole office in the background.

Hearing aids

A software layer that works with your hardware — not instead of it.

Sound Lab

Hear the difference

Exempel 1

Mix
Voice

Exempel 2

Mix
Voice

Exempel 3

Mix
Voice

30s commercial · 2nd place in Commercial of the Year 2024/25

Global Goals

Linked Global Goals

Voice Choice is more than a tool — it is the right to choose who you want to hear.

3

Good Health and Well-Being

Sound is an invisible factor that affects our health more than we think. Whether it is constant background noise in a classroom, the stress of a hectic work environment, or the tangible exhaustion that noise pollution creates in urban settings, it affects our well-being. For people with neurodevelopmental disabilities, our app can be the difference between a productive day and an overwhelming experience filled with auditory fatigue and anxiety. We offer a solution where the user can eliminate distracting sounds without isolating themselves entirely from their surroundings. In the longer term, this helps reduce the risk of noise-related health problems such as high blood pressure, sleep difficulties and mental exhaustion.

4

Quality Education

School environments are often filled with sound — students talking, chairs scraping, and surrounding noise. For students with ADHD or autism, this can be an insurmountable obstacle. With Voice Choice, students can listen selectively to the teacher and filter out distracting sounds, creating a more inclusive educational environment and giving more people the chance to succeed academically.

10

Reduced Inequalities

People with a reduced ability to filter sound are today forced to adapt to a world that is not designed for their needs, which can lead to social isolation. Our technology is an individualisation measure that evens out differences. Our pricing model is designed not to exclude anyone, by offering a free version so that everyone can access the technology regardless of financial circumstances.

Pricing

Simple: free to start.

Free
0 kr

forever

  • One voice profile
  • Basic noise filtering
  • Full privacy — everything on-device
Join the beta
Premium
100 kr/mån

at launch

  • Unlimited voice profiles
  • Environment profiles: classroom, office, restaurant
  • Customizable EQ
  • Full engine access
Join the beta

Open beta is planned on Google Play in 2026 — free for everyone on the waitlist. Pricing applies from launch.

About

From the school cafeteria to the national championship

Autumn 2024. The cafeteria at Lugnetgymnasiet in Falun. Fifty conversations competing for the same air. Your friends opposite you are talking, but you don't hear a word — not because you aren't listening, but because your brain can't sort it. That was the situation I was in when the thought struck me: "what if I could choose to hear only my friends and no one else".

That thought is how Voice Choice was born, and the friends sitting opposite me were Klas and Arvid. I shared the idea and how I wanted them with me as co-founders. They loved it and ran Voice Choice UF with me. But after the UF year, Klas and Arvid chose not to continue, so I carried it forward on my own.

I, Milton Rohde Dahlberg — CEO, ML architect, app developer, UI/UX designer, writer. 20 years old. Today I carry the entire technical product myself — AI training, Android architecture, DSP implementation, design. I work as an orchestrator for AI models and autonomous coding agents, an approach that didn't exist five years ago.

Voice Choice UF won 2nd place in Commercial of the Year in Dalarna 2024/25 and Innovation of the Year Dalarna 2024/25, which qualified us for the national championship. In 2026, after the UF year, I won 2nd prize in the jagharenide.nu idea competition. A clear vision and desire drives me on: to give everyone the right to own their own sound world.

Innovation of the Year Commercial of the Year 2nd prize — jagharenide.nu 2026 SM 2025
Milton Rohde Dahlberg — VD och ML-arkitekt för Voice Choice

Contact

Say hello.

Questions, partnerships, or early testing — we read everything.

Email voicechoiceuf@gmail.com Instagram @voicechoiceuf YouTube @VoiceChoiceUF

FAQ

Good to know.

How does Voice Choice isolate specific voices?
Voice Choice uses an on-device AI model that analyzes a 10-second voice sample to create a unique voice profile. The AI then filters ambient audio in real time, isolating that specific voice while suppressing background noise. All processing happens locally on your Android device.
Is my voice data sent to the cloud?
No. All AI processing happens entirely on your device. Voice profiles are stored locally, and no audio data ever leaves your phone — this ensures absolute privacy and eliminates network latency.
What do I need to use Voice Choice?
You need an Android device and a pair of headphones. For the full experience, noise-canceling headphones work best — Voice Choice layers on top of your existing hardware, letting specific voices through while your headphones block the rest.
When will Voice Choice be available?
Open beta is planned for Google Play in 2026. Join our waitlist to be notified when we launch.
How much does it cost?
Voice Choice uses a freemium model. The free tier includes one voice profile and basic filtering. Premium at 100 SEK per month unlocks unlimited voice profiles, environment profiling, customizable EQ, and full engine access.
Who is Voice Choice for?
Voice Choice is built for anyone who struggles to hear specific voices in noisy environments — particularly students in classrooms, people with ADHD, autism, or sensory sensitivities, and professionals in open offices.

The waitlist

Be first when the beta opens.

Open beta on Google Play, 2026. One email when it opens — then we go quiet.

We store your email address only. No spam, no sharing — ever.