Neural network-based speech recognition technology for voice assistants and IoT devices

WhisPro™ is an AVS qualified neural network based keyword spotting (KWS) software package, allowing customers to add wake-words and commands to voice-enabled IoT devices. As a voice user interface, WhisPro is targeting the rapidly growing use of voice as a primary human interface for intelligent cloud-based services and edge devices.

WhisPro provides a customizable (any language, any keyword) wake-word and keyword voice interface. This allows users of headsets, wearables, smart TVs, smart speakers, smartphones , infotainment systems, and other voice-enabled IoT devices to interact with cloud-based voice assistant services, as well as enabling voice activation on edge devices through its small footprint and processing needs, without cloud connectivity.


WhisPro voice user interface enables wake words and commands for always listening devices

High recognition rate with very low false acceptance rate operating on-device
Scalable RNN speech AI technology supports simultaneous multi wake-word phrases
Built-in ambient noise cancellation engine for improved user experience in noisy environments

Main Features

  • Customizable wake words and commands service to enable voice control for a range of use cases and end markets
  • Speaker independent, pronunciation agnostic, and support for multiple languages
  • Amazon Voice Services (AVS) qualified wake word solution, ready for Alexa enabled devices
  • Improved performance when combined with a voice front end software, like Ceva’s ClearVox
  • The WhisPro SDKit provides guidelines and tools to integrate and test the WhisPro technology in a target system
  • Low power (<10MIPS) and low memory footprint (<75KB) model on Ceva-BX2 DSP, that can also run on resource constrained MCUs as small as Cortex M0+
  • Speaker verification algorithms identify unique users through voice biometrics
  • WhisPro voice user interface is available for Ceva’s DSPs: Ceva-BX1, Ceva-BX2 and SensPro2 family, as well as Arm CPUs

Block Diagram

WhisPro Diagram

Ceva WhisPro Text2Model Voice User Interface

Instant Keyword Generation Demo

Ceva WhisPro voice User Interface Text2Model is the future of voice commands technology. It’s a unique zero data, fully customizable, keyword spotting solution that runs on resource-constrained hardware. Using a companion app, device OEMs or the end user can define and generate wake words to detect any phrase using simple text input, without internet connection. Text2Model handles any phrase that can be transcribed by Latin letters.