FAQs
We understand that integrating with the SpeechAce platform involves a lot of moving parts. To help you get started, we've put together some key information that we hope you’ll find useful!
Last updated
We understand that integrating with the SpeechAce platform involves a lot of moving parts. To help you get started, we've put together some key information that we hope you’ll find useful!
Last updated
Most speech recognition providers focus on transcribing what the user is speaking. Speechace is designed and trained to assess how well the user is speaking. Our focus is on evaluating the spoken language of a non-native speaker – i.e. how intelligible, natural and rich is their language. Our models are trained and optimized for this specific purpose.
The Speechace API is a technology building block which allows you to add speaking activities with AI scoring within any eLearning platform. If you are looking for a ready to use solution, please use the Contact form below and our team will share additional options less technical than the API.
All of our API plans come with a free trial. You can click ‘Start Free Trial’ on any of the and get started.
Yes, we offer volume, bulk, and discounted annual plans.
Please use the on our website and share details about your use case and volume. Our team will share details.
If your platform and developers are ready, and you follow the docs and samples, you can implement one speaking activity in less than 1 week and trial it with users. Our API trial is designed to allow you to test out the API before you have to pay for a subscription.
We go to great lengths in running independent, blind, randomized evaluations for all our models to ensure our API results are within close agreement range with qualified human evaluators. Our API has been independently verified by both Academic and Corporate entities.
Speechace tackles the challenge of accent through ensuring the training and validation datasets represent a wide range of speakers in terms of demographics, language proficiency and accent. This allows our general model to be optimized for the target result of measuring intelligible, natural, and fluent speech irrespective of accent.
The mere presence of accent is not considered a criteria for penalizing a speaker, so long as it does not interfere with intelligibility and fluency.
By default, CORS (Cross-Origin Resource Sharing) is disabled on the Speechace API to protect the security of your API key. This means that direct calls from the frontend, whether your application is running locally (e.g., http://localhost) or in a live production environment (e.g., https://yourwebsite.com), are blocked by the browser due to security reasons.
To securely access the Speechace API, it should be called from your backend server. If you'd still like to make API calls directly from the frontend, we can enable CORS on your subscription, but please note that this exposes your API key to the client side and introduces security risks.
If you'd still like to enable CORS, please contact Speechace support at contact@speechace.com.
Please fill out the and we’d be glad to share technical reports detailing our performance and accuracy metrics.