Software Development Kit for On-device Speech Recognition

iOS, Android, and custom platforms

SDK Documentation

Why on-device?

Why rely on 3rd-party cloud services, or build and scale your own ASR backend, when on-device speech recognition is viable for a number of use cases.

Works Offline

No internet conectivity is required

Privacy (COPPA) and Security

Speech data does not leave the device

Predictable Pricing

No per-use fees that don't align with your business model


Language and acoustic models can be customized to specific domain and population


On-device processing

iOS, Android
Unity plugin

Real-time processing

30k words on iPhone 6, over 100k words on iPhone 7

Deep Neural Networks

State-of-the-art engine based on Deep Neural Networks


Natural Language Understanding extracts user’s intent (coming soon)

We offer professional services for SDK integration, proof-of-concept development, customization of language and acoustic models, and porting to custom hardware platforms.

Use cases


Spoken language instruction and assessment, early reading, early childhood games


Add voice interaction to your mobile games via unity plugin


Extend your virtual and augmented reality apps with voice interaction


Simplify UX with voice search

Voice control in apps

Circumvent complex navigation

Industrial and Automotive

Hands-free voice-only control


Dashboard provides developers with the ability to seamlessly transfer audio recordings and speech recognition metadata from their apps to the cloud, during the app development phase

Debug, evaluate, and tune your apps and gain insights about speech recognition performance via Dashboard



KeenASR SDK provides simple yet flexible API

KeenASR SDK Documentation

Review SDK documentation and a simple proof-of-concept iOS and Android apps on Github.