SoundHound Teaches 22 Languages to Houndify Voice AI
Houndify provides a platform for all or part of the voice AI for companies that operate around the world. Cars built by Hyundai and Kia, White Castle restaurants, streaming services like Pandora, and social media apps like Snapchat all rely on Houndify to provide voice interactions. Houndify supports an automatic speech recognition (ASR) and natural language understanding (NLU) engine to converse with users and respond appropriately. The customization options for developers include the expected common languages like Spanish, German, Japanese, and even English with an Indian accent, as well as some less common languages such as Russian and Korean. Notably, Houndify can converse in Hebrew, which neither Alexa nor Google Assistant can do. Apps and devices can include the list of language options or default to whichever tongue is most widely spoken in the area. This is just the first round of new languages, according to SoundHound. The long-term plan will see more than 100 languages among Houndify’s options eventually.
“We’re committed to our efforts in language expansion, including key ASR features such as long-form transcription and customized vocabularies,” SoundHound CEO Keyvan Mohajer said. “Our ASR adds tremendous value and leveraging our breakthrough NLU provides endless possibilities.”
Houndify’s popularity has grown out of its ability to mimic human conversation. It supports voice assistants capable of understanding complicated requests using casual vocabulary, even when there are multiple commands in a single sentence. Augmenting that capability with a lot more languages will be a boon for plenty of SoundHound clients. For instance, SoundHound set up a partnership with parking services provider Parkopedia last year to include Parkopedia’s data in Houndify as a third-party service for automotive clients. As Parkopedia’s list covers more than 70 million parking spaces in 15,000 cities and 89 countries, the extra languages may draw more users. Similarly, hotels using SoundHound internationally may appreciate the linguistic options.
“We’ve always been focused on providing a voice AI platform that allows our partners to speak the languages of their users,” SoundHound vice president of engineering Majid Emami said in a statement. “This announcement reinforces the growth of our language library, ready for any company looking for speech recognition with global reach.”