Generate captions - mcut docs

Provider packages produce TranscriptResult. The core transcription package can turn that result into caption elements and subtitle files.

import { buildApplyCaptionsCommand, toCaptionElements, toSrt } from "@mcut/transcription"

const elements = toCaptionElements(transcript, {
  maxChars: 36,
  maxGapMs: 800,
})

const command = buildApplyCaptionsCommand(elements, {
  trackName: "Captions",
})

engine.dispatch(command)

const srt = toSrt(transcript)

Use server-side providers for API-key services. Use the local Whisper provider only in browsers that pass its WebGPU capability checks.