YouTube Transcript Scraper – Captions, Subtitles & Video Data

YouTube transcript scraper: get the full transcript, captions or subtitles of any public YouTube video, with timestamps, in any available language (human-made or auto-generated), plus video title, channel, views, duration and publish date. No login or API key.

Price: $2.00 per 1,000 transcripts. Videos without captions, private or removed videos and blocked requests produce a free error row and are never charged.

Use this instead of a general web crawler or a browser-based YouTube scraper when you need the spoken text of videos, fast and in a clean format for LLMs, RAG, search or analysis.

What data do you get?

• 📝 Full transcript as one plain-text field (transcript), ready for LLMs and NLP
• ⏱ Timestamped segments: start time, duration and text of every caption line (segments)
• 🌍 Language info: language code and name of the returned captions, whether they are auto-generated (speech recognition) or human-made, and the list of all caption languages the video has
• 🎬 Video metadata: title, channel name, channel ID and URL, publish date, duration, view count, category, keywords (tags), description (e-mail addresses and phone numbers removed), thumbnail URL, live-stream flag
• 🔢 Counts: number of caption lines, words and characters

Works with normal videos, Shorts, past live streams, music videos, lectures and podcasts, including multi-hour videos (we tested a 4.5-hour course: 50,000+ words in one item).

Why use this YouTube transcript scraper?

• Reliable: in our test runs on Apify, 100% of the captioned videos returned a transcript (65 of 65, none ended blocked). The Actor rotates residential IPs automatically when YouTube shows a bot check, so you don't need to set up a proxy.
• Fast: plain HTTP requests, no browser. About 1 second per video, 5 videos in parallel.
• Fair: you pay only for transcripts you receive. Errors are free.
• No login, no API key, no YouTube quota.
• Clean output: flat fields, HTML entities decoded, formatting tags removed.

Use cases

• AI and LLM pipelines: summarize videos, build RAG knowledge bases from lectures, podcasts and tutorials, or create training and evaluation data.
• Content repurposing: turn videos into blog posts, newsletters, show notes, social posts or subtitles.
• SEO and marketing research: see what competitors say in their videos and which keywords they use.
• Market and brand monitoring: search the spoken content of product reviews, earnings calls and news videos for brand or product mentions.
• Education and accessibility: get readable text versions of lessons and talks.
• Academic research: build text datasets of public video content for linguistics, media and political science studies.

How to get YouTube transcripts

1. Click Try for free / Start.
2. Paste one or more video links into YouTube videos (https://www.youtube.com/watch?v=..., https://youtu.be/..., /shorts/... or just the 11-character video ID).
3. Optionally choose a Transcript language (e.g. en, de, es) and Caption type.
4. Click Start and download your data as JSON, CSV, Excel, XML or HTML, or connect it to Make, Zapier, n8n, Google Sheets, webhooks or the Apify API.

Input

Field	Type	Description
videos	array of strings	Video URLs or IDs. Watch, youtu.be, Shorts, live, embed and mobile links all work. Channel and playlist links are not supported.
language	string	Optional language code (en, de, pt-BR, ...). Empty = the video's original spoken language.
captionType	any / manual / auto	any (default) prefers human-made captions and falls back to auto-generated ones.
languageFallback	boolean	If the video has no captions in language, return the original-language transcript (default true). If false, you get a free error row instead.
includeTimestamps	boolean	Add the segments array with start time and duration of each line (default true).
includeMetadata	boolean	Add publish date, duration, views, category, keywords, description and thumbnail (default true).
proxyConfiguration	object	Leave off. The Actor handles proxies itself.

Example input:

{
    "videos": [
        "https://www.youtube.com/watch?v=iG9CE55wbtY",
        "https://youtu.be/arj7oStGLkU",
        "UF8uR6Z6KLc"
    ],
    "language": "en",
    "captionType": "any",
    "includeTimestamps": true,
    "includeMetadata": true
}

Output

Each dataset item has a type: transcript or error. Real output from a run on Apify (shortened; the real item has all 427 segments and 65 caption languages):

{
    "type": "transcript",
    "videoId": "iG9CE55wbtY",
    "url": "https://www.youtube.com/watch?v=iG9CE55wbtY",
    "title": "Do schools kill creativity? | Sir Ken Robinson | TED",
    "channelName": "TED",
    "channelId": "UCAuUUnT6oDeKwE6v1NGQxug",
    "channelUrl": "https://www.youtube.com/channel/UCAuUUnT6oDeKwE6v1NGQxug",
    "publishDate": "2007-01-06T21:00:34-08:00",
    "durationSeconds": 1203,
    "viewCount": 25025865,
    "category": "Education",
    "keywords": ["Ken", "Robinson", "TED", "TEDTalks"],
    "description": "Sir Ken Robinson makes an entertaining and profoundly moving case for creating an education system that nurtures creativity ...",
    "thumbnailUrl": "https://i.ytimg.com/vi/iG9CE55wbtY/sddefault.jpg",
    "isLiveContent": false,
    "language": "en",
    "languageName": "English",
    "isAutoGenerated": false,
    "requestedLanguageMissing": false,
    "availableLanguages": [
        { "languageCode": "af", "languageName": "Afrikaans", "isAutoGenerated": false },
        { "languageCode": "sq", "languageName": "Albanian", "isAutoGenerated": false },
        { "languageCode": "ar", "languageName": "Arabic", "isAutoGenerated": false }
    ],
    "segmentCount": 427,
    "wordCount": 3170,
    "characterCount": 17574,
    "transcript": "Good morning. How are you? (Audience) Good. It's been great, hasn't it? I've been blown away by the whole thing. In fact, I'm leaving. (Laughter) ...",
    "segments": [
        { "start": 27.103, "duration": 2.575, "text": "Good morning. How are you?" },
        { "start": 29.702, "duration": 1.403, "text": "(Audience) Good." },
        { "start": 31.129, "duration": 1.668, "text": "It's been great, hasn't it?" }
    ],
    "scrapedAt": "2026-10-08T16:18:54.178Z"
}

Error row (free):

{
    "type": "error",
    "videoId": "LXb3EKWsInQ",
    "url": "https://www.youtube.com/watch?v=LXb3EKWsInQ",
    "title": "COSTA RICA IN 4K 60fps HDR (ULTRA HD)",
    "errorCode": "NO_CAPTIONS",
    "error": "This video has no captions or transcript.",
    "availableLanguages": []
}

Possible errorCode values: NO_CAPTIONS, LANGUAGE_NOT_AVAILABLE, VIDEO_UNAVAILABLE, PRIVATE_VIDEO, AGE_RESTRICTED, MEMBERS_ONLY, LIVE_NOT_STARTED, BLOCKED, CAPTIONS_FAILED, INVALID_INPUT, SCRAPE_FAILED.

Times in segments are in seconds. The dataset has two ready-made views in Apify Console: Overview and Video metadata.

Pricing

This Actor uses pay-per-event pricing. You pay only for transcripts you get. Platform usage and proxies are included.

Event	Price	When it is charged
transcript	$0.002 per video ($2.00 / 1,000)	Each transcript saved to the dataset
Errors	Free	No captions, unavailable/private/age-restricted videos, blocked requests, invalid input

Examples:

Job	Cost
10 videos	$0.02
1,000 videos	$2.00
10,000 videos	$20.00

Long videos cost the same as short ones. Tip: set Max cost per run in the run options. The scraper stops cleanly when your budget is reached.

Use with AI agents (MCP)

Add the Actor as a tool in the Apify MCP server (https://mcp.apify.com?tools=openharvest/youtube-transcript-scraper), or let agents find it with search-actors ("youtube transcript").

Example call-actor input for "Summarize this YouTube video for me":

{ "videos": ["https://www.youtube.com/watch?v=UF8uR6Z6KLc"], "includeTimestamps": false, "includeMetadata": false }

Example input for "Get the German subtitles of these two talks, with timestamps":

{ "videos": ["https://youtu.be/arj7oStGLkU", "https://youtu.be/iG9CE55wbtY"], "language": "de", "languageFallback": false }

Tip for agents: includeTimestamps: false keeps the output small (just transcript), which saves context tokens.

FAQ

Which videos work? Any public or unlisted video that has captions: human-made subtitles or YouTube's auto-generated captions. Most spoken videos in common languages have auto-generated captions. Pure music or nature videos often have none; you get a free NO_CAPTIONS row.

Can I get the transcript in another language? Yes, if the video has captions in that language (see availableLanguages). Machine translation of captions is not supported: YouTube blocks translated caption requests without a logged-in browser session. If your language is missing, you get the original-language transcript (requestedLanguageMissing: true) or, with languageFallback: false, a free error row.

Does it work for Shorts and live streams? Shorts: yes. Live streams: yes, once the stream has ended and YouTube has processed the captions. Upcoming streams return LIVE_NOT_STARTED.

Can I scrape a whole channel or playlist? Not in this Actor. Enter the individual video URLs.

Can I get age-restricted, private or members-only videos? No. These need a logged-in account, and this Actor never logs in. They return a free error row.

What does BLOCKED mean? YouTube showed its "confirm you're not a bot" check on every IP the Actor tried for that video. This is rare (0 of 65 videos in our tests) and temporary. Run the video again later. You are not charged.

Do I need a proxy? No. On Apify the Actor uses Apify residential proxy automatically, because YouTube blocks datacenter IPs. The cost is included in the price.

Can I run it via API or on a schedule? Yes. Use the Apify API, the JavaScript/Python clients, Schedules, or integrations such as Make, Zapier, n8n and Google Sheets.

Is it legal to scrape YouTube transcripts?

The Actor reads only public data: the captions and basic video details that YouTube shows to any visitor without login. It does not log in, does not use cookies of real accounts, does not download video or audio, and does not collect personal data: e-mail addresses and phone numbers in video descriptions are replaced with [email removed] / [phone removed]. The channel name is included only as the public author of the video. Transcripts are copyrighted by their creators. Respect their rights, YouTube's Terms of Service and laws such as GDPR when you store, publish or train on the data. Quoting, research and analysis are usually fine; republishing full transcripts may not be. Do not use the data for spam or unsolicited outreach. If you are unsure, consult a lawyer.

Related actors

By the same developer:

• Apple Podcasts Scraper – Episodes, Charts & Podcast RSS Feeds
• Bilibili Scraper – Videos, Search, Rankings, Comments & Danmaku
• Keyword Suggestions – Google, YouTube & Amazon Autocomplete
• PDF to Markdown – DOCX, PPTX, XLSX & HTML for LLM & RAG
• Substack Scraper – Newsletter Posts, Full Text & Comments
• Telegram Channel Scraper – Posts, Messages & Views, No Login

Support & feedback

Found a bug or missing field? Open an issue in the Issues tab. We usually respond within a day. Feature requests are welcome.