Speak + OpenAI Speech Recognition

SAN FRANCISCO, CA – Speak today announced it is a launch partner for OpenAI’s new automatic speech recognition API, which builds on earlier “Whisper” models the team previously open-sourced. Highly performant speech recognition is an important driver of Speak’s product experience, which helps users learn new languages through a focus on speaking and listening.

Speak has already begun testing the API in production and is seeing meaningful performance impact through reduced latency and higher word recognition accuracy. Whisper represents a step-function improvement in speech recognition, enabling more open-ended conversational experiences that weren’t previously possible. Whisper’s performance on accented speech as well as seamless handling of multi-lingual speech (code-switching) makes it even more powerful in applications related to language learning. With this new technology, Speak can provide more accurate, real-time feedback to its users, which will help them learn new languages faster.

As a launch partner, Speak will continue to partner closely with OpenAI on new technology launches, and will apply these tools to expand access to effective language learning. Speak believes that the use of innovative technologies such as automatic speech recognition will revolutionize language learning, and is committed to being at the forefront of this revolution.

For more information about Speak, please visit speak.com.