On-device Speech Recognition for Kids Mobile Apps

Create fun and engaging voice experiences in education and entertainment apps

KeenASR SDK is a software development kit for on-device automatic speech recognition on mobile devices and custom hardware platforms. The SDK runs on iOS, Android, ChromeOS, and Linux operating systems.

KeenASR SDK offers a number of advantages over cloud-based ASR services for use in mobile apps and custom hardware products focused on children. Whether it's an educational or entertainment app, KeenASR SDK can provide highly accurate automatic speech recognition in your product.

The SDK is designed with privacy in mind; since all the processing happens locally, speech data doesn't need to leave the device, thus providing COPPA compliance by design.

Key Features

On-Device Processing

On-device Processing

COPPA compliant; works offline; real-time results, without latency; no need to scale your backend to address peaks in usage.

Models optimized for kids voices

Models optimized for children's voices

Our acoustic models have been optimized to work for kids voices and in noisy environments.

Oral reading instruction and assessment

Oral Reading Instruction and Assessment

SDK provides features optimized for oral reading instruction & assessment, and pronunciation assessment.

Cross platform support

Cross Platform Support

iOS, Android, ChromeOS, and Linux. Grow your customer base and increase revenue for voice-based products that run on all relevant platforms.

Development Tools

Developer Friendly

Simple, easy to integrate, API; Unity support; easy to evaluate via trial SDK; cloud based tools for development support; ability to programmatically define what the recognizer is listening to at any given time.

Relevant Use Cases

KeenASR SDK can be effectively used in following scenarios:

  • Oral reading instruction and assessment
  • Second language learning and assessment
  • Early childhood (3-6) learning; vocabulary, numeracy, phonics
  • Speech impediment therapy and assessment
  • Interactive videos
  • Simple dialogs with characters
  • Spoken interaction in fictional languages (Klingon, Na'vi, Elvish, and similar)
  • Voice-based app navigation (command and control)
  • Content search

Hardware and System Requirements

KeenASR SDK provides on-device automatic speech recognition functionality for devices running iOS, Android, ChromeOS, and Linux operating systems. The SDK is linked against your mobile application and requires access to the ASR Bundle for each language your app needs to support. ASR Bundles are typically included in your app, but they could also be downloaded the first time your application starts.

  • OS: iOS 11 or later. Android 5.0 or later.
  • CPU: A8 (iPhone 6 or newer; equivalent iPads, and Android devices)
  • RAM: ~100MB or more
  • Storage: ~30MB of on-device storage for the ASR Bundle

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