YouTube Transcript Scraper

Extract timed transcripts and captions from YouTube videos in bulk, with no API key, no OAuth and no quota. Give it a list of videos and it returns one row per caption line, with the start time, the duration, the text and the language code.

What you can do with it

• Feed video transcripts to an LLM for summarising, tagging or question answering
• Build a searchable archive of a channel's spoken content
• Pull subtitles for translation or repurposing into articles and social posts
• Analyse how competitors talk about a topic across hundreds of videos
• Time-align quotes back to the exact second they were said

Input

Field	Type	Notes
videoIds	array	Bare video IDs, watch URLs, youtu.be links or /shorts/ links. Mixed input is fine.
language	string	Optional two-letter code such as en or ja. Falls back to the video's first available track.
proxyConfiguration	object	Recommended. YouTube refuses most datacenter addresses.

{
  "videoIds": [
    "dQw4w9WgXcQ",
    "https://youtu.be/9bZkp7q19f0",
    "https://www.youtube.com/shorts/kX0KTNBiisM"
  ],
  "language": "en",
  "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Output

One dataset item per caption line:

{
  "video_id": "dQw4w9WgXcQ",
  "start": 18.64,
  "duration": 3.24,
  "text": "We're no strangers to love",
  "language": "en"
}

Export as JSON, CSV, Excel or XML from the dataset, or read it straight from the Apify API.

Languages and auto-generated captions

Both human-written and auto-generated (ASR) tracks are supported, in every language YouTube offers them. Set language to prefer a specific track; when that language is not available the video's first track is used instead, and the language field on each row tells you which one you actually got.

Pricing

Pay per result. You are charged for each transcript line delivered, plus a small per-run start fee. Videos that are skipped cost you nothing beyond that start fee.

Videos that return nothing

Some videos genuinely have no transcript, and this actor tells you which rather than failing the whole run:

• No captions available — the uploader disabled them and no auto-generated track exists
• Age-restricted or private — the video cannot be read without signing in
• Region-blocked — not available from the address the run used

Every skipped video is reported in the run log with its reason, and the run finishes normally with the transcripts it did get.

Why a proxy is recommended

YouTube refuses most datacenter addresses with a bot check. With RESIDENTIAL proxy groups enabled the check does not fire, and runs complete without retries. Without a proxy, expect a portion of your videos to be skipped.

FAQ

Do I need a YouTube API key? No. There is no key, no OAuth flow and no daily quota.

How many videos can I pass at once? There is no fixed limit. The run processes them in order and reports what it could not read.

Can I get the whole transcript as one block of text? Each row is one caption line with its timing. Concatenate the text fields in order to get a single block, or keep the timings to link back to the moment in the video.

Does it work on Shorts? Yes, when the Short has captions. Pass the /shorts/ URL directly.