VideoUltra API
    Preparing search index...

    Function createCaptionsEndpoint

    • A Transcriber for your own captions endpoint (DECISIONS #164, #167): any speech service, any language, with the key on your server. Sends the sound in pieces of up to 30 s, cut where it's quietest (no word cut in two; silent pieces aren't sent), up to three at once (one while the service is busy), and turns the replies into timed words: words said in the silence before speech move onto it, made-up and doubtful words are dropped. Detect decides the language from the first pieces with speech and asks for it from then on. When the service's own word times stop well before a piece ends while people still speak, the rest of the piece is asked for once more.

      import { createCaptionsEndpoint } from '@video-ultra/ai';
      <VideoEditor transcriber={createCaptionsEndpoint({ endpoint: '/api/captions' })} />

      Parameters

      Returns Transcriber