Elsa Speech Assessment API

 

Overview

The ELSA API empowers developers to build AI-driven applications that assess the accuracy and effectiveness of spoken English communication. Using advanced machine learning technology, the API provides a simple and flexible interface for integrating speech evaluation into your application.

The API supports two types of speech analysis:

• Scripted speech (with transcript)
• Unscripted speech (without transcript)

For both types, ELSA evaluates key aspects of spoken English, including:

• Pronunciation
• Prosody (intonation and rhythm)
• Fluency

For unscripted recordings, the API additionally evaluates:

• Grammar
• Vocabulary

Detailed feedback is provided across these dimensions to help users identify mistakes and improve their speaking skills.

Scripted API Updates

The latest version of the Scripted API offers several improvements and extended features. Refer to the Scripted API section of this documentation for a full overview.