Voiceitt’s automatic speech recognition (ASR) enables users with non-standard speech to access and interact with your product by voice. This real-time speech recognition technology is suitable for a variety of contexts, including conversation, dictation and video call captioning. The Voiceitt API lets you offer an accessible voice solution and expand your offering to more users by enabling speech-to-text for users with non-standard speech. This guide will get you started.
The base URL of the Voiceitt HTTP API is https://api2.voiceitt.com/ .
Authentication
The Voiceitt HTTP and WebSockets API are based upon authenticated sessions which are tracked using JSON Web Tokens (JWT). An authenticated JWT token is obtained by providing valid credentials to either the login/user_id or login/email endpoint.
The authenticated JWT token is subsequently passed:
- in the HTTP header of each call to the transcribe endpoint or
- at the initialisation of a WebSockets connection.
Speaker-independent and personalised speech-to-text
Voiceitt's speech-to-text API can be used to access speech-to-text services which are either:
- not personalised to the particular speaker (speaker-independent)
- personalised to the particular speaker (also called speaker-dependent)*
The authentication method used determines whether speaker-independent or personalised speech-to-text is invoked in subsequent API calls. More specifically use:
- login/user_id for speaker-independent speech-to-text.
- login/email for personalised speech-to-text.
*Personalised speech-to-text typically delivers accuracy improvements over speaker-independent speech-to-text. To access personalised speech-to-text for a particular user, the user must have registered with the Voiceitt application https://web.voiceitt.com/ and have completed the enrolment process.