YouTube & Bandcamp Lyrics Transcriber

Extract approximate, timestamped lyric transcripts from user-provided YouTube videos, direct audio/video URLs, or Bandcamp track/album pages.

Provide one or more audio/video pages and the Actor returns structured transcription results with timestamps, source metadata, warnings, and billing information.

Input

{
  "startUrls": [
    { "url": "https://www.youtube.com/watch?v=..." },
    { "url": "https://artist.bandcamp.com/album/...", "trackTitle": "Track Name" }
  ],
  "quality": "balanced",
  "language": "en",
  "audioFilter": false,
  "maxDurationSeconds": 900,
  "includeSegments": true,
  "includeFullText": true
}

Input fields

Field	Default	What it does
startUrls	required	List of audio/video pages to transcribe. Supports YouTube video URLs, direct audio/video URLs, Bandcamp track URLs, and Bandcamp album URLs. For Bandcamp albums, add trackTitle to the individual item so the Actor selects the intended track instead of guessing.
quality	balanced	Transcription quality/speed tradeoff. fast is quicker and cheaper for previews. balanced is recommended for most songs. accurate can improve difficult vocals but takes longer. multilingual is recommended for non-English or mixed-language audio.
language	en	Optional language hint, e.g. en or de. Leave empty if you want automatic language detection with multilingual mode.
maxItems	5	Maximum number of input URLs processed from startUrls.
maxDurationSeconds	900	Safety limit for one audio item. Audio longer than this is skipped with an error. Set 0 to disable the limit, but long files cost more and take longer.
audioFilter	false	Enables an optional voice/silence filter. For songs, false often preserves repeated sung lines better; for speech-heavy audio, true can reduce silence/noise.
includeSegments	true	Include timestamped segments (start, end, text) in the dataset output. Disable if you only need metadata/full text.
includeFullText	true	Include joined transcript text as transcriptText. Disable if you only want timestamped segments or want smaller output items.

Output

Each dataset item contains:

• url, sourceType, title, artist, duration
• billableAudioMinutes: audio duration rounded up to full minutes for pay-per-event billing
• segments: timestamped transcript segments
• transcriptText: joined text when enabled
• transcriptMeta: language, duration, quality mode, and filter settings used for the run
• warnings and error fields when extraction/transcription is partial

Pricing

This Actor is priced per started audio minute. A 4:12 track counts as 5 billable minutes. Failed items are not charged.

Notes

• Music vocals are hard for automated transcription. Treat output as approximate, especially for electronic, processed, layered, or reverberated vocals.
• Some video pages may block automated access or require user/session verification; those URLs will return an error instead of silently fabricating output.
• For Bandcamp album pages with several tracks, pass trackTitle in the startUrl item to select the intended track.