With the release of VoiceHub and the companion apps for Android and iOS, the company has integrated support for TrulyNatural, its large vocabulary speech recognition solution that runs on device with customizable natural language understanding (NLU) capabilities. Voice user interface (VUI) designers, says the company, can leverage this new capability to produce natural language-enabled products capable of supporting dozens of languages and dialects - ideal for creating products targeting a global market.
"Providing free access to flexible tools like VoiceHub helps to accelerate the next wave of branded voice experiences and domain-specific, customized voice assistants," says Todd Mozer, CEO at Sensory. "Since releasing the VoiceHub beta in October, VUI designers around the world have used it to create hundreds of voice AI models for dozens of automotive, wearable, smart speaker and smart home products. Our new large vocabulary and NLU capabilities will unlock enhanced VUI functionality with intents and entities. This empowers designers to develop more complex and truly conversational voice interfaces with the capability of handling millions of unique phrases."
The company says it has been providing early access to select partners and the response has been overwhelmingly positive.
For example, Daniel Colonna, Microcontrollers Marketing Director at STMicroelectronics says, "By adding VoiceHub to the menu of Sensory technologies accessible to our STM32 Discovery Board customers, developers are able to design working prototypes of natural language-capable products in a manner of minutes, not days, dramatically reducing their time to market."
On January 27th, the two companies will jointly host a VoiceHub webinar focused on creating large vocabulary NLU models for the STM32 Discovery kit. Attendees will learn how to create custom wake words, domain-specific language models, and integrate with the ST development platforms. The companies are working together on rapid domain-specific voice UI prototyping and development, where designers can have a test file ready within an hour of building their model.
For wake word phrase spotted commands, says the company,