YouTube Transcript & Subtitle Extractor

Extract transcripts, subtitles, and captions from any YouTube video — standard videos, Shorts, and live streams. Get timestamped segments, full text, and metadata in JSON, SRT, or plain text.

Features

• Auto-generated & manual captions — extracts existing YouTube subtitles first (fastest, highest quality).
• Speech-to-text fallback — when no captions exist, downloads audio and transcribes via Deepgram Nova-2.
• 50+ languages — supports any language available through YouTube captions or STT.
• Multiple output formats — JSON (with timestamps), plain text, or SRT subtitles.
• Video metadata — title, author, duration, thumbnail, and upload date included.
• No login required — works with any public YouTube video.

Use Cases

• Content repurposing — turn YouTube videos into blog posts, articles, or social media content.
• Accessibility — generate transcripts for hearing-impaired audiences.
• Research & analysis — build datasets of video content for market research or NLP tasks.
• SEO — extract text content from videos for search engine optimisation.
• Translation — get transcripts as a starting point for multilingual subtitles.
• Education — create study materials from lecture videos.

Input

Field	Type	Default	Description
url	string	(required)	YouTube video URL
language	string	"en"	Preferred transcript language (ISO 639-1)
format	string	"json"	Output format: "json", "text", or "srt"
includeMetadata	boolean	true	Include video metadata in the output

Supported URL Formats

• https://www.youtube.com/watch?v=VIDEO_ID — standard videos
• https://m.youtube.com/watch?v=VIDEO_ID — mobile URLs
• https://youtube.com/watch?v=VIDEO_ID — bare domain
• https://youtu.be/VIDEO_ID — short URLs
• https://www.youtube.com/shorts/VIDEO_ID — YouTube Shorts
• https://www.youtube.com/live/VIDEO_ID — live streams

Output

JSON format (default)

{
    "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "platform": "youtube",
    "metadata": {
        "title": "Rick Astley - Never Gonna Give You Up (Official Music Video)",
        "author": "Rick Astley",
        "duration_seconds": 212,
        "thumbnail_url": "https://i.ytimg.com/vi/dQw4w9WgXcQ/maxresdefault.jpg",
        "upload_date": "20091025"
    },
    "transcript": {
        "language": "en",
        "segments": [
            {
                "start": 0.0,
                "end": 3.5,
                "text": "We're no strangers to love"
            },
            {
                "start": 3.5,
                "end": 7.2,
                "text": "You know the rules and so do I"
            }
        ],
        "full_text": "We're no strangers to love You know the rules and so do I ..."
    }
}

Error output

{
    "error": "NO_TRANSCRIPT",
    "message": "No transcript available for this video: https://...",
    "url": "https://..."
}

Error Codes

Code	Description
INVALID_URL	The provided URL is not a valid YouTube URL.
NO_TRANSCRIPT	The video has no captions and STT transcription failed.
VIDEO_NOT_FOUND	The video was not found, is private, or has been removed.
RATE_LIMITED	Too many requests — the actor uses proxy escalation to mitigate this.
PLATFORM_ERROR	An unexpected error occurred during extraction.
STT_FAILED	Speech-to-text transcription failed (audio download or API error).

How It Works

1. URL validation — verifies the input is a recognised YouTube URL format.
2. Subtitle download — uses yt-dlp to fetch auto-generated or manual captions (JSON3 > VTT > SRT preference).
3. STT fallback — if no captions exist, downloads the audio track and sends it to Deepgram Nova-2.
4. LLM cleaning — optionally cleans STT output (punctuation, filler words) via an LLM.
5. Proxy escalation — starts with datacenter proxies, escalates to residential if rate-limited.

FAQ

Does this require a YouTube login? No. The actor works with any public YouTube video without authentication.

What about auto-generated captions vs. manual subtitles? The actor fetches both and uses whichever is available. Most YouTube videos have auto-generated captions, making the STT fallback rarely needed.

Which languages are supported? Any language supported by YouTube's caption system (100+ languages for auto-generated, unlimited for manual). The STT fallback supports 30+ languages via Deepgram.

Can I extract from YouTube Shorts? Yes. YouTube Shorts URLs (youtube.com/shorts/VIDEO_ID) are fully supported.

Can I extract from live streams? Yes, for completed live streams with replay available. Currently live streams in progress are not supported.

What happens if a video has no captions and STT fails? The actor returns an error with code NO_TRANSCRIPT and a descriptive message.