# Changelog of YouTube Transcript Scraper — Bulk Captions & Subtitles (`vero-api/youtube-transcript-scraper`) Actor

- **URL**: https://apify.com/vero-api/youtube-transcript-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/vero-api/youtube-transcript-scraper.md

## Changelog

### 2026-10-25

- **Much faster runs.** Up to 20 videos are transcribed at a time, and a single video comes back in about a second.
- Fewer videos come back as "couldn't be loaded right now".
- Rows arrive as each video finishes, so they can be in a different order from your list. Every row carries its `url` and `videoId`.
- **Auto-detect returns the language spoken in the video.** It used to follow YouTube's default captions for English-speaking viewers, so a Japanese video could come back in English.
- **Type any YouTube caption code** in Transcript language, such as `iw`, `zh-Hant`, `pt-BR` or `es-419`. Hebrew and Filipino now work, with either `iw`/`he` or `fil`/`tl`, and `zh-TW`, `zh-HK` and `zh-CN` pick Traditional or Simplified Chinese.
- Human-made captions now win over auto-generated ones even when they carry a regional code, so `en` (or Auto-detect) gets human-made `en-GB` captions instead of the auto-generated English track.
- **New: Fallback languages.** Codes to try in order when a video lacks your first choice; add `auto` to end with the spoken language.
- **New: Caption type.** Accept any captions, only human-made ones, or only auto-generated ones.
- Word counts are now real word counts for Japanese, Chinese, Thai and other languages written without spaces.
- Auto-generated SRT and WebVTT lines no longer overlap the next line. Human-made subtitles keep their original timing.
- Styled captions (outlined or colored text) no longer repeat every line twice in the transcript.
- Rows without a transcript show their reason in the Output table's **Note** column.
- The run summary now counts links that aren't single videos and links left at your spending limit.

### 2026-10-24

- **Many videos per run.** The input is now `videoUrls`, a list of links or video IDs, replacing the single `videoUrl`.
- Lower price: $2 per 1,000 transcripts.
- A video without a transcript no longer stops the run: it gets its own free row explaining why.