ELSA Releases API for English Tutor Voice AI
Educational app developer ELSA (English Language Speech Assistant) has released an API of its proprietary AI, suitable for integrating into other linguistic educational tech platforms. The company has partnered with Brazilian, Indian, and Japanese educational technology companies to extend access to Elsa’s voice far beyond the ELSA Speak App and its more than 25 million global users.
The Elsa iOS and Android mobile apps offer a voice assistant to guide those learning or looking to improve their English. The app listens to how the user pronounces sentences broken down by individual sounds, diagnosing where they are making mistakes and giving real-time suggestions on how to adjust their speech. The AI is trained on Elsa’s large database of voices in various non-native accents and can spot subtle mistakes in rhythm, intonation, and other subtle language distinctions. The API embeds that AI in other services. Elsa has teamed with edtech companies around the world including the Japanese corporate education firm Edulinx, Brazilian English center chain CNA Go and Indian UOLO and staffing platform Sahi.ai, among others. The startup raised $15 million about a year ago, with that funding likely contributing to the API development.
“As English communication quickly evolves into a requirement for the modern skillset, it’s essential that edtech and learning companies include this competency in their curriculum,” Elsa CEO Vu Van explained. “Students want digitized training and quality learning at their fingertips anytime, anywhere and the release of our API makes it available to a broader audience.”
Educational voice AI is becoming more common as the technology improves and educators learn more about how it can augment their efforts. The field saw a spike of interest as the COVID-19 pandemic led to more at-home learning too. Largely that has been centered around children, but much of the new services could be used by adults learning a new language too. Amazon introduced Alexa voice profiles for kids with the Reading Sidekick AI tutor. Google has been very active as well, with a new teaching toolkit for Google Play Books, an Android app called Read Along with its kid-focused Diya voice assistant, and folding reading tutor app Rivet into Google Assistant. Even Duolingo, which counts many adults as well as children among its users, has pumped resources into the voice AI element. The app developer built a text-to-speech tool to provide the characters who illustrate its lessons with unique voices. Each cartoon character will get a synthetic voice to match the fictional background and personality Duolingo has devised as part of an overall upgrade to the app’s AI and natural language understanding.
Elsa’s API plan also shares similarities with a deal that textbook publisher McGraw Hill recently inked with children’s speech recognition tech startup SoapBox Labs to integrate voice tech into its services. It’s one of several deals SoapBox has made in recent months to integrate voice AI into reading education. The startup has built a new literacy assessment for kids with edtech software developer Amplify and the SoapBox Fluency tool for automatically tracking performance for teachers.