YouTube Transcript Scraper (No API Key, No Login)

Get the transcript of any public YouTube video as one row per video, from video links, whole channels, playlists or searches. Each row has the plain text, timestamped segments, optional SRT or WebVTT subtitles and RAG chunks, plus the video's title, channel, publish date, views and caption languages. You don't need a login, an API key or cookies.

What people use it for:

• Search or chat over video: RAG chunks carry start and end times, so an answer can link to the right moment.
• Reusing talks, podcasts and tutorials as text for articles, show notes and summaries.
• Researching a channel or topic: every transcript from a channel's latest videos or a YouTube search, in one table.
• Subtitle files: SRT and WebVTT in your language, from human-made or auto-generated captions.
• Watching channels: schedule it with onlyNew to get only the new videos' transcripts.

Quick start

Paste this into the JSON tab of the input and click Start:

{"urls": ["https://www.youtube.com/watch?v=arj7oStGLkU", "https://youtu.be/dQw4w9WgXcQ", "https://www.youtube.com/watch?v=UF8uR6Z6KLc"], "languages": ["en"]}

That returns 3 transcripts in about 10 seconds. For a channel, paste its link or @handle, for example https://www.youtube.com/@TED with "maxVideosPerChannel": 5. Add "chunkSize": 300, "chunkOverlap": 40 for RAG chunks.

What you get

Group	Fields
Video	videoId, url, title, channelName, channelHandle, channelId, channelUrl, publishDate (YYYY-MM-DD, the date shown on the video's page), publishedAt (exact time, UTC), durationSeconds, duration, viewCount, likeCount, category, thumbnailUrl
Transcript	hasTranscript, transcript (plain text), wordCount, characterCount, segmentCount
Formats	segments (start, end, text; on by default), srt, vtt (when asked), chunks (index, start, end, wordCount, text; when chunkSize is set)
Language	language, languageName, isAutoGenerated, isTranslated, translatedFrom, isFallbackLanguage, availableLanguages (every caption track the video offers), captionLanguages (every language code with captions, once each), manualCaptionLanguages (codes with human-made captions, not auto-generated)
No transcript	noTranscriptReason (no-captions, no-matching-language, private, age-restricted, members-only, live-now, upcoming, temporarily-blocked, translation-refused and others), noTranscriptMessage
Source	sourceType (video, channel, playlist, search), sourceInput, sourcePosition, scrapedAt

Inputs can be video links (watch, youtu.be, Shorts, live, embed) or bare IDs, channel links or @handles (the latest N videos), playlist links, and YouTube searches (the top N results in YouTube's order). You can mix them in one run, and each video is read once.

One real row (run on 2026-10-18, with the text, segments and the language lists cut short):

{
  "videoId": "arj7oStGLkU",
  "url": "https://www.youtube.com/watch?v=arj7oStGLkU",
  "title": "Inside the Mind of a Master Procrastinator | Tim Urban | TED",
  "channelName": "TED",
  "channelHandle": "@TED",
  "channelId": "UCAuUUnT6oDeKwE6v1NGQxug",
  "publishedAt": "2016-04-06T16:59:35.000Z",
  "publishDate": "2016-04-06",
  "durationSeconds": 844,
  "duration": "14:04",
  "viewCount": 62072457,
  "likeCount": 2055195,
  "category": "People & Blogs",
  "hasTranscript": true,
  "language": "en",
  "languageName": "English",
  "isAutoGenerated": false,
  "isTranslated": false,
  "isFallbackLanguage": false,
  "availableLanguages": [{"code": "ar", "name": "Arabic", "isAutoGenerated": false}, {"code": "bg", "name": "Bulgarian", "isAutoGenerated": false}],
  "captionLanguages": ["ar", "bg"],
  "manualCaptionLanguages": ["ar", "bg"],
  "transcript": "So in college, I was a government major, which means I had to write a lot of papers. ...",
  "wordCount": 2277,
  "characterCount": 12671,
  "segmentCount": 315,
  "segments": [{"start": 12.645, "end": 14.015, "text": "So in college,"}, {"start": 15.349, "end": 16.913, "text": "I was a government major,"}],
  "sourceType": "video",
  "sourcePosition": null
}

A video without a transcript is a free row with hasTranscript: false, a noTranscriptReason and a plain-words noTranscriptMessage. The dataset has three views (Overview, Video details, No transcript (free)) and exports to JSON, CSV, Excel or HTML. The run's SUMMARY record counts videos by outcome and lists anything that failed.

Price

You pay per transcript returned. There is no charge for Apify platform usage on top.

Row	Free and Starter	Scale	Business
Transcript	$1.50 per 1,000	$1.125 per 1,000	$0.825 per 1,000

Plus $0.00005 per run start. Apify's free $5 monthly credit covers about 3,300 transcripts. Free-plan runs stop at 200 videos each; any paid Apify plan removes the cap. Example: a channel's latest 50 videos cost $0.075.

Free rows: videos with no captions (or none in your languages), private, removed, age-restricted and members-only videos, live streams and premieres that haven't finished, videos YouTube refused to show, videos skipped by onlyNew, and the same video found twice in one run (read once). A run that returns nothing because of the input costs nothing. To cap a run's spend, set maxVideos or Maximum cost per run in the run options.

Limits

• It returns the captions YouTube has: human-made, auto-generated or YouTube's automatic translation. It doesn't transcribe audio, so a video with no captions at all comes back as a free row. Auto-generated captions have no punctuation in some languages and can mishear names.
• List languages in order of preference (en, es, pt-BR, ...). Human-made captions win over auto-generated ones. When a video lacks your language, you get the video's own language (isFallbackLanguage: true, on by default), YouTube's translation (translate, off by default) or a free row saying so.
• YouTube refuses translated captions far more often than the captions themselves. When it refuses, you get the video's own language if the fallback is on, or a free translation-refused row if it is off.
• It sees what a logged-out viewer sees. Age-restricted, members-only and private videos need a signed-in viewer, and this Actor never signs in, so they come back as free rows.
• YouTube sometimes refuses a request. The Actor then retries from a new IP address, moving from datacenter to residential proxies only when it has to. A video YouTube keeps refusing becomes a free temporarily-blocked row, and you can run it again later.
• Sizes: up to 5,000 latest videos per channel (10 by default; Shorts and past live streams only with includeShorts), up to 5,000 per playlist (100 by default) and up to 500 results per search (10 by default).
• Views and likes are YouTube's public counts at the time of the run. Likes are missing when the channel hides them.

Only new since last run, and schedules

Put a channel, playlist or search in the input, turn on onlyNew and schedule the Actor under Schedules in Apify Console. Each run then returns and charges for only the videos this input (same sources and language settings) hasn't returned before. Videos that came back without a transcript are not remembered, so they are tried again next time.

Integrations: API, n8n, Make and AI agents

API. Run it and get the rows back in one call:

curl -X POST "https://api.apify.com/v2/acts/locaihost~youtube-transcripts/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://www.youtube.com/watch?v=arj7oStGLkU"], "languages": ["en"], "chunkSize": 300}'

Apify's API clients for Python and JavaScript work the same way.

n8n, Make, Zapier, Google Sheets, Slack, webhooks. Apify's integrations can send each run's dataset on, for example new transcripts from a channel into a vector database or a Google Doc.

AI agents (MCP). Add this Actor to Claude, Cursor or any MCP client with https://mcp.apify.com/?tools=locaihost/youtube-transcripts, then ask something like "Summarise what the last 5 videos on this channel say about pricing."

FAQ

How much does it cost to scrape YouTube transcripts?

$1.50 per 1,000 transcripts on the Free and Starter plans, less on Scale and Business. Videos without captions are free. Apify's free $5 monthly credit covers about 3,300 transcripts.

Do I need a YouTube API key?

No. The official YouTube Data API doesn't return transcripts of other people's videos. This Actor reads the captions a logged-out viewer can see.

Can I use it through the API or from Python?

Yes. Use the run-sync-get-dataset-items call above, or Apify's Python and JavaScript clients. Every run's data is also available as JSON, CSV or Excel from its dataset.

Can I use it from an AI agent?

Yes. Through the Apify MCP server (https://mcp.apify.com/?tools=locaihost/youtube-transcripts) any MCP client can call it with video, channel or playlist links and read the transcripts back.

Is it legal to scrape YouTube transcripts?

The Actor reads the captions and public video details that YouTube shows a viewer who isn't signed in. It never signs in and never reads private, members-only or age-restricted videos. Transcripts are the creators' work and stay under their copyright: use them for analysis, research, search, accessibility and internal tools, not to republish as your own content. YouTube's terms restrict automated access, so check that your use fits them and the laws where you work. You are responsible for how you use the results.

Why are some fields empty?

A video without a transcript has transcript: null and says why in noTranscriptReason. likeCount is empty when the channel hides likes, srt and vtt appear only when you turn them on, and chunks only when chunkSize is above 0.

Which languages work?

Any language YouTube has captions for. Use codes like en, es, de, ja, pt-BR. Leave languages empty to get each video's own language.

Can I see the publish date and which caption languages a video has?

Yes, on every row, including the free rows for videos without a transcript. publishDate is the date shown on the video's page (2024-03-01) and publishedAt is the exact moment in UTC, so near midnight the two can name different days. captionLanguages lists every language with captions (["en", "es", "fr"]) and manualCaptionLanguages lists the ones made by people. That lets you pick out videos with human-made English captions before you pay for transcripts. Both come from the same requests as the transcript and cost nothing extra.

More from locaihost data

• Doc to Markdown: turn PDFs and office documents into clean Markdown and chunks for the same RAG pipeline.
• OCR for scanned documents: text from scans and images when there's no text layer to read.
• TikTok Trends Scraper: trending hashtags and videos by country, to find topics worth covering.
• Google Trends Scraper: check whether a topic is growing before you make a video about it.

Did it help?

If this saved you time, a quick rating on the Store helps other people find it.

Support

Something not working, or a field you need? Open an issue on the Issues tab: we answer issues within 24 hours.