NABILA · Deep learning for personalized hearing, on the devices you ownComing soon
01 / NABILA

The next step in audio starts with the listener.

Meet NABILA: a deep learning model we're building to support hearing-aid users inside the apps they already use. Because hearing is different for everyone.

02 / NABILA

Your hearing is the starting point.

The same starting point audiology has always had: a hearing test maps how you hear at every frequency. That chart is your audiogram.

03 / NABILA

First, meet the recipe: NAL-R.

Hearing-aid fitting has relied on trusted prescription formulas for decades. The best known is NAL-R: read each threshold on your audiogram, and prescribe back roughly half of the missing hearing, frequency by frequency. Everything NABILA does starts from this idea: your profile and the sound in front of you.

04 / NABILA

More than turning up the volume.

Every frequency matters. NABILA uses the hearing profile to adjust frequency levels, instead of turning everything up together.

THE SAME AUDIO
Listener 01
Listener 02
Different needs. Different processing.

Same audio · different hearing needs

Hearing profileAUDIOGRAM
2505001k2k4k6k

Six thresholds guide the transformation

Personalization starts with your hearing.

Your audiogram is the starting point

Incoming audio
Your audiogram
NABILAHearing-profile-conditioned processing
Audio shaped for your profile

Audiogram in. Prescription levels out.

EVERY FREQUENCY COUNTS01 → 06

Levels informed by the audiogram

Personalize frequencies, not just volume.

Different frequencies. Tailored levels.

ONE SOUND. DIFFERENT WAYS TO HEAR.

The detail that changes everything? You.

The same audio goes in. The hearing profile guides the change. Switch profiles to explore the idea: personalization is more than louder sound.

The same incoming audio
NABILA+Hearing profile A

A different emphasis on higher frequencies.

A visual exploration of personalization
NABILA / APISDKs coming soon
01 / INTEGRATE

Your favorite tools. Sound for everyone.

From web to mobile, connect audio to your listener’s hearing profile. NABILA libraries bring integration to the languages and stacks you work with.

NABILASDK
phpPHP
TSTypeScript
PyPython

One idea. In your language.

LaravelNext.jsNode.jsDjangoFastAPI
Part of the upcoming release
02 / STREAM

The conversation continues. Sound adapts.

Streaming sends small audio chunks for processing against the hearing profile and returns them in sequence. No waiting for the call to end or a complete file to upload.

Example: a call sent in audio chunksSTREAM

Incoming chunks from your app

0102030405060708
↓NABILA↓

Chunks returned after processing

0102030405060708

Scroll to follow each chunk through the journey.

03 / TRUST

Privacy is part of the design.

Our service design follows one principle: your data serves your request only. Process audio to return a result, without retaining content or using it to train the model.

01

For your request

Temporary processing of audio and hearing profiles to produce the requested result.

02

No retention

No audio or hearing-profile storage after the request is a principle of the service design.

03

Not for training

Request content stays outside model training and development data.

Keep scrolling. There’s more to the story.01 / 03
Coming soon · Early access

Try NABILA before the public launch.

Leave your email and tell us how you would use NABILA — for yourself, or for an app that supports people with hearing aids.

Your invitation arrives by email before the public launch.