Made in Norway

Separate any audio in real time

AI models for on-device processing.

audio

100%

speech

noise

30%

alarm

reverb

50%

music

Running fully on-device

audio

100%

speech

noise

30%

alarm

reverb

50%

music

Trusted by industry leaders

From Hollywood to Formula 1, our tech already runs in the world's toughest environments.

Listen for yourself.

This audio is processed locally on your device. Our WebAssembly uses your CPU for realtime audio separation.

One audio stream. Many controllable components

Decide what gets captured. Control what you hear. Shape the audio experience for yourself, other humans, and voice agents.

Built to run on-device

10 ms latency

Designed for real-time audio processing on-device. Also available at 21, 32, or 96 ms.

Only 241 kb

Runs on 40% of a single Cortex‑A53 core. From embedded hardware to cloud backends.

Secure by design

Audio never leaves the device, and all processing runs locally. Privacy is preserved for all users.

Shit in, shit clean out.

AI models for any audio challenge

Isolate voices and clean up dialogue without losing the human in the signal. Built for critical communications, conferencing, accessibility, and any product where every word has to land clearly.

One SDK, every platform.

Linux

Windows

macOS

iOS

Android

Embedded

Embed in minutes.
Ship today.

Drop the Rapidly SDK into any stack.

Docs

main.py

import rapidlyengine = rapidly.RapidlyEngine()engine.add_license("lk_...")  # valid license key removes watermarkprocessor = engine.create_processor(    "models/speech-denoise-32ms.v1.0.rapidly", 2, 44100)# audio_block: float32 ndarray, shape (num_samples, num_channels)processed = processor.process(audio_block)

30-day trial

Test in your stack.

Free 30-day trial license, no watermark.

Used where audio happens.

iOS Dictation App

Cleaner input and higher accuracy for your speech-to-text pipeline. No noise, reverb, rustle or wind.

Clear Communication

Embedded into radios, headsets, and wearables. Remove noise while voices and critical sounds pass through, like alarms, sirens, drones, and whistles.

Desktop Conferencing

Add studio-grade clarity to any macOS conferencing app. Rapidly boosts speech intelligibility and removes keyboard clicks, fan noise and room reverb.

Live Streaming

Give streamers enhanced audio straight from the browser. Rapidly ships as WebAssembly, so noise, room reverb, and other unwanted sounds get cleaned in real time on device.

Smart Speakers

Turn any smart speaker into a karaoke machine. One tap strips vocals from any song so anyone in the room becomes the lead singer.

Competitive Gaming

Give competitive players an audio edge. Isolate footsteps and reloads from chatter and explosions so critical cues come through clean.