YouTube Transcripts Scraper (Alt Instances) (Apify Actor)

Scrape transcripts for specific YouTube videos using a fallback chain:

1. optional customTranscriptUrlTemplate
2. Invidious captions (/api/v1/captions/{video_id}) if enabled, using a built-in large instance list
3. YouTube internal player API fallback (android/ios clients)
4. YouTube watch-page caption tracks fallback
5. optional youtube-transcript-api fallback
6. youtubetranscript.com if enabled (default on)
7. optional yt-dlp fallback (last resort, disabled by default; auto-disabled on Apify when no cookies are provided)
8. automatic retry pass for transient blocking/errors

Input

Use .actor/input_schema.json in the Apify UI. altInstances is internal and not exposed as user input. For best reliability on Apify, keep useApifyProxy=true. When proxyGroups is empty, the Actor now tries RESIDENTIAL first (if available) and falls back automatically. Increase proxyPoolSize (default 8) to rotate more sessions when YouTube blocks individual IPs. Keep fetchVideoMeta=false for faster transcript-only runs. If yt-dlp still hits bot-check, provide cookies via youtubeCookies input or YOUTUBE_COOKIES env var (Apify secret env var is recommended).

Output

Dataset items: one transcript record per input video.

Key fields:

• video_id, url, title
• status (found or missing)
• language, source, transcript_url
• text, word_count, error

Key-value store:

• OUTPUT: run metadata + totals + warnings (no item list, rely on dataset)
• For misses, warnings may include missing_transcript_debug ... with per-source error trail.
• Apify log now prints run_totals and (when missing) missing_item video_id=... source=... error=....

Notes

• Transcript availability depends on captions being available and endpoint accessibility.
• This Actor does not use the official YouTube Data API.