Frequently Asked Questions

FAQ

What is the minimum hardware requirement?

• CPU: ARM Cortex-A7 or equivalent (400MHz+)

• RAM: 5MB free (for models + runtime)

• Storage: 10MB (for binaries + models)

• OS: Linux (any distribution with pthreads)

How accurate is the detection?

See our benchmarks for detailed accuracy metrics. In typical conditions, accuracy exceeds 93% across all sound types, with false alarm rates under 3%.

Can I use it on a microcontroller (MCU)?

Currently we support Linux-based embedded systems. MCU support (FreeRTOS, bare metal) is on our roadmap. Contact us if you have specific requirements.

How does licensing work?

Each device that runs the SDK requires a license key. Licenses are per-device with volume-based pricing. Contact us for a quote.

Is network access required?

No. The SDK runs entirely offline. All inference happens locally on the device. A network connection is never needed during operation.

Can you add support for a new sound type?

Yes. Contact us with your requirements for custom model development.