# Changelog of YouTube Transcript & Content Extractor (`fixitfoundry/youtube-transcript-actor`) Actor

- **URL**: https://apify.com/fixitfoundry/youtube-transcript-actor/changelog.md
- **Full Actor documentation**: https://apify.com/fixitfoundry/youtube-transcript-actor.md

## Changelog

### v0.2 - 2026-07-06

#### Fixed

- **CRASH BUG**: `YouTubeTranscriptApi.getTranscript()` was renamed to `fetchTranscript()` in youtube-transcript v1.2.1. The actor crashed on startup with `TypeError: Cannot read properties of undefined`. This was the root cause of the QA test failure.
- **Proxy routing**: Added Apify proxy integration via Crawlee's `gotScraping`. All YouTube API calls (transcript via InnerTube, page scraping for metadata) now route through the `BUYPROXIES94952` proxy group. This fixes transcript and metadata fetch failures from Apify's datacenter IPs.
- **Dataset schema**: Made all optional fields accept `null` values (`type: ["number", "null"]` instead of `type: "number"`) to prevent schema validation errors when page scraping fails gracefully.
- **CAPTCHA detection**: Added explicit `g-recaptcha` check in page scraping to log a clean warning instead of a cryptic error when YouTube serves a captcha wall.

#### Changed

- Page scraping now uses proxy-routed requests (same proxy as transcript) for consistent behavior.
- Transcript error handling: downgraded from `console.error` to `console.warn` — missing transcript is a common edge case, not an actor failure.