Pure audio transcription

We are very excited to announce that today we are launching the POST /dictation API, our solution to transcribe an audio file or stream. For users who are familiar with our POST /speech endpoint, this API is similar in that it provides the audio transcript, but does not extract any intents, entities, nor traits.

This new API supports up to 5 minutes of audio to allow for transcribing live audio with multiple sentences. In order to reduce latency, this API is ideal for any implementations that have been using POST /speech so far, but do not need the intents, entities, and traits information extracted. We encourage strict speech-to-text use cases to transition to POST /dictation today. Same requests, faster responses and longer audio inputs.

Feel free to contact us with any questions or comments.

Thanks for using Wit.ai as we continue to improve!

Wit.ai Team