Keyword spotting
Voice commands get lost in the noise.
Your product has to hear a short command over a busy room. We clean the voice with AI speech denoising before keyword spotting, and help you build it in.
Last updated
Request a demo
How it works
From a noisy order to a recognised command
A café is loud and keywords are short. Denoising cleans the voice first, so keyword spotting hears the words, not the room.
Your system: Sound
A spoken command in a busy room.
Your system: Microphone
The microphone in your product.
Our solution: Speech denoising
Removes the noise around the voice.
Our solution: Keyword spotting
Recognises the words you choose.
Your system: Coffee machine
Starts the drink that was ordered.
What it handles
Commands, not chatter
We tune it to your words and your setting, then check each detection against what was really said in the recording.
What it catches
- Your keywords
- The words your product needs, such as drink names.
- Different voices
- High and low, fast and slow.
- Short commands
- A single word, said once.
- Busy moments
- Commands spoken over a full room.
- Your microphone
- The microphone in your product.
What it ignores
- Silence
- Pauses and empty rooms.
- Other words
- Speech that isn't one of your keywords.
- Machine noise
- Grinders and fans.
- Steam
- Milk frothing and steam wands.
- Clatter
- Cups, cutlery and doors.
Applications
Where it's built in
Your product gets the keyword it heard, as it's spoken.

Coffee machines
Drink orders spoken at the machine.

Vending machines
Hands-free choices at self-service machines.

Kitchen appliances
Voice control while hands are busy.
Evaluation
Interested in this solution?
Get in touch with our engineering team to evaluate it for your system.
Evaluation platform. Firmware and software are under development; each use case requires application-specific validation.
Other use cases
More on denoising
See how we clean up speech for calls.
Speech denoising
Removes background noise from speech as it's spoken, so calls and recordings stay clear.

