YouTube Transcript Scraper — Extract Captions, Subtitles & Full Video Text

The fastest way to extract YouTube transcripts, captions, and subtitles at scale. Get the complete spoken text from any YouTube video as structured, timestamped data — with full video metadata included in every result.

---

Why Use This YouTube Transcript Scraper?

• 3,000+ videos per minute — the fastest YouTube transcript extractor on Apify
• No browser, no API key, no login required — fully anonymous, zero setup
• Bulk processing — feed 1, 100, or 10,000+ video URLs in a single run
• Rich metadata — title, author, channel, views, duration, publish date, category, description, and keywords on every result
• Timestamped segments — each caption segment includes start time and duration for subtitle generation
• Full transcript text — all segments joined into one searchable plain text block
• 20+ languages — request transcripts in English, Spanish, German, French, Japanese, Korean, Chinese, Portuguese, and more
• Smart language fallback — always returns something, even if your preferred language isn't available
• Auto-generated captions supported — works with both manual and YouTube's auto-generated subtitles
• YouTube Shorts supported — extract transcripts from Shorts the same way as regular videos
• Migration-safe — if Apify moves the container mid-run, it resumes exactly where it left off

---

What Data Do You Get?

Every result includes two layers of data — the transcript itself and the video metadata — so you never need a separate scraper for context.

Transcript Data

Field	Description
transcript_text	The entire spoken content joined into one plain text string — ready for NLP, indexing, or content repurposing
segments	Array of timestamped caption segments, each with start (seconds), dur (duration), and text
segment_count	Total number of caption segments in the transcript
status	Result status: ok, no_captions, unplayable, or error

Video Metadata

Field	Description
title	Video title
author	Channel name
channel_url	Direct link to the channel
view_count	Total views at time of scraping
duration_seconds	Video length in seconds
publish_date	Original upload date (ISO 8601)
category	YouTube category (Music, Education, Entertainment, etc.)
description	Full video description text
keywords	SEO keywords/tags set by the uploader
is_family_safe	Whether the video is marked as family-friendly

---

Input

{
  "videoIds": [
    "https://www.youtube.com/watch?v=jNQXAC9IVRw",
    "dQw4w9WgXcQ",
    "https://youtu.be/9bZkp7q19f0"
  ],
  "maxVideos": 1000
}

Parameter	Type	Required	Description
videoIds	string[]	Yes	YouTube video URLs or bare 11-character IDs. Accepts watch URLs, youtu.be short links, and Shorts URLs.
maxVideos	integer	No	Maximum videos to process in this run. Default: unlimited for paid plans.

Accepted URL Formats

All of these work — paste whatever you have:

https://www.youtube.com/watch?v=jNQXAC9IVRw
https://youtu.be/jNQXAC9IVRw
https://www.youtube.com/shorts/jNQXAC9IVRw
jNQXAC9IVRw

---

Sample Output

{
  "video_id": "jNQXAC9IVRw",
  "video_url": "https://www.youtube.com/watch?v=jNQXAC9IVRw",
  "title": "Me at the zoo",
  "author": "jawed",
  "channel_id": "UC4QobU6STFB0P71PMvOGN5A",
  "channel_url": "https://www.youtube.com/@jaboreda",
  "view_count": 386567321,
  "duration_seconds": 19,
  "publish_date": "2005-04-23T20:31:52-07:00",
  "category": "Film & Animation",
  "description": "The first video on YouTube.",
  "keywords": ["me at the zoo", "jawed karim"],
  "is_family_safe": true,
  "is_private": false,
  "transcript_text": "All right so here we are in front of the elephants the cool thing about these guys is that they have really really really long trunks and that's cool and that's pretty much all there is to say",
  "segment_count": 3,
  "segments": [
    { "start": 1.0, "dur": 5.0, "text": "All right so here we are in front of the elephants" },
    { "start": 6.0, "dur": 8.0, "text": "the cool thing about these guys is that they have really really really long trunks" },
    { "start": 14.0, "dur": 5.0, "text": "and that's cool and that's pretty much all there is to say" }
  ],
  "status": "ok"
}

---

How Fast Is It?

Scenario	Videos	Time
Single video	1	~2 seconds
Small batch (10 videos)	10	~5 seconds
Medium batch (100 videos)	100	~15 seconds
Large batch (1,000 videos)	1,000	~30 seconds
Bulk run (5,000 videos)	5,000	~3 minutes

Processing is fully parallel — five videos are fetched simultaneously, each with independent metadata and transcript retrieval. A 1,000-video run with all transcripts available completes in under 30 seconds.

---

Use Cases

AI Training Data & Machine Learning

YouTube is the largest source of spoken language on the internet. This scraper turns that into structured text you can feed directly into ML pipelines:

• Build speech-to-text training datasets from millions of hours of captioned video
• Create question-answering datasets from educational and tutorial content
• Extract text corpora for NLP model training — sentiment analysis, topic modeling, summarization
• Build domain-specific knowledge bases from lecture series, conference talks, and expert interviews
• Fine-tune large language models on YouTube content in any language or domain
• Generate training pairs by combining transcript text with video metadata (title, category, keywords)

Content Marketing & SEO

Turn competitor videos into actionable content intelligence:

• Analyze competitor video scripts — see exactly what they say and how they structure their content
• Extract keywords and topics from top-ranking YouTube videos in your niche
• Repurpose video content into blog posts, articles, social media threads, and email newsletters
• Identify content gaps by analyzing what topics your competitors cover that you don't
• Build searchable video archives — index transcript text for internal content libraries
• Monitor brand mentions — search transcripts for your brand, product, or competitor names

Academic Research & Journalism

• Track public statements from politicians, executives, and public figures across thousands of videos
• Analyze media narratives by extracting and comparing transcripts from news channels
• Study language patterns — accent analysis, vocabulary usage, discourse structures
• Build longitudinal datasets — track how messaging changes over time across channels
• Archive spoken content — create permanent text records of video content that may be deleted

Accessibility & Localization

• Generate subtitle files for videos that lack them — use the timestamped segments to create SRT or VTT files
• Build translation pipelines — extract source language transcripts, translate, and sync back to video timing
• Create searchable archives for hearing-impaired audiences
• Audit caption quality — compare auto-generated captions against manual transcriptions

Business Intelligence & Competitive Analysis

• Monitor earnings calls and investor presentations uploaded to YouTube
• Track product launch announcements from competitors
• Analyze conference talks and keynotes for industry trends
• Extract customer testimonial data from review and unboxing videos
• Monitor thought leadership content across your industry

Podcast & Video Production

• Generate show notes from podcast episodes hosted on YouTube
• Create searchable episode archives for podcast back-catalogs
• Extract quotes and highlights for social media promotion
• Build episode summaries by feeding transcripts into AI summarization tools
• Cross-reference content across episodes to find recurring topics

---

Supported Languages

The scraper supports any language that YouTube provides captions for. These are the most commonly available:

Code	Language	Code	Language
en	English	ja	Japanese
en-US	English (US)	ko	Korean
en-GB	English (UK)	hi	Hindi
es	Spanish	id	Indonesian
es-ES	Spanish (Spain)	fil	Filipino
de	German	vi	Vietnamese
fr	French	tr	Turkish
it	Italian	ru	Russian
pt	Portuguese	zh	Chinese (Simplified)
pt-BR	Portuguese (Brazil)	zh-Hant	Chinese (Traditional)

If the video has captions in a language not listed above, enter the language code manually — the scraper will attempt to find and return it.

---

Status Values Explained

Every result includes a status field so you know exactly what happened:

Status	What It Means
ok	Transcript extracted successfully
no_captions	Video exists but has no caption tracks — common for music videos, very old uploads, and some regional content
unplayable	Video is private, deleted, age-restricted, or region-locked
error	Unexpected failure — the video ID was processed but something went wrong. Will be retried if you re-run with the same input.

---

How to Get Video IDs in Bulk

This scraper needs video IDs or URLs as input. Here are the best ways to collect them:

1. YouTube Search Scraper — search for keywords and collect all matching video IDs
2. YouTube Channel Scraper — get every video ID from a specific channel
3. YouTube Playlist URLs — extract IDs from playlist pages
4. Google Sheets — paste a column of YouTube URLs and export as a list
5. Manual collection — copy-paste video URLs directly from your browser

Feed the collected IDs into this scraper and get transcripts for all of them in one run.

---

Output Formats

Results are available in all Apify-supported formats directly from the dataset:

• JSON — structured data, best for programmatic use and API integration
• CSV — flat table format, opens directly in Excel and Google Sheets
• Excel (.xlsx) — native spreadsheet format with proper column headers
• XML — structured markup for enterprise integrations
• RSS — feed format for monitoring and alerting workflows

---

Integration & Automation

Connect to Your Workflow

Apify integrations let you automatically send transcript data to:

• Google Sheets — append new transcripts to a spreadsheet in real-time
• Slack / Discord — get notified when transcripts are ready
• Webhooks — trigger your own API when a run completes
• Zapier / Make — connect to 5,000+ apps without code
• Amazon S3 / Google Cloud Storage — store transcript archives in the cloud
• BigQuery / Snowflake — load transcripts into your data warehouse

API Access

Start runs programmatically via the Apify API:

curl -X POST "https://api.apify.com/v2/acts/YOUR_ACTOR_ID/runs" \
  -H "Content-Type: application/json" \
  -d '{"videoIds": ["dQw4w9WgXcQ", "jNQXAC9IVRw"]}'

Scheduling

Set up recurring runs to monitor new video transcripts daily, weekly, or at any custom interval using Apify's built-in scheduler.

---

Frequently Asked Questions

Can I extract transcripts from any YouTube video? Only from videos that have captions enabled. Most videos uploaded after 2020 have auto-generated English captions. Older videos, music-only videos, and some regional content may have no captions — these are returned with status: no_captions so you can filter them.

Does this work with YouTube's auto-generated captions? Yes. Auto-generated captions are returned identically to manually uploaded ones. The text may be less accurate for auto-generated captions (especially for technical jargon, names, or accented speech) but is still highly usable for NLP, indexing, and content analysis.

What happens if my requested language isn't available? The scraper never returns empty just because one language is missing. It falls back automatically — first to English, then to whatever language track is available. The response always tells you exactly which language was used via the lang and lang_fallback fields.

Can I process thousands of videos in one run? Yes. There is no hard limit on input size. The scraper processes videos in parallel batches of 5 with automatic rate limiting. A 1,000-video run takes about 30 seconds. A 10,000-video run takes about 5 minutes.

How do I convert the output to SRT subtitle files? The segments array contains start (seconds) and dur (duration) for each text block. Use any SRT converter library to transform these into standard subtitle format. The timing data is already in the right structure.

What is transcript_text vs segments? transcript_text is the entire transcript joined into one plain text string — use this for NLP processing, keyword extraction, search indexing, or feeding into AI models. segments is the array of individual timestamped caption blocks — use this for subtitle generation, time-aligned analysis, or building video navigation.

Does this scraper work on private or age-restricted videos? Private and deleted videos return status: unplayable. Age-restricted videos also return unplayable since they require authentication.

Does the scraper handle Apify container migrations? Yes. Every completed video ID is saved to persistent storage every 60 seconds and on every migration event. When the run resumes on a new container, it skips already-completed videos and continues from exactly where it stopped. Zero data is lost or duplicated.

Can I use this for AI and LLM training data? Yes. The output is designed for downstream AI use. transcript_text gives you clean text for training. segments gives you time-aligned data for speech models. Combine with metadata fields (title, category, keywords, description) to build richly labeled datasets.

Is a proxy required? No. This scraper works without any proxy on Apify's standard infrastructure. For extremely high volumes (50,000+ videos per day), you can optionally enable residential proxy rotation in the input settings — but it is not required for normal use.

Can I extract transcripts from YouTube Shorts? Yes. YouTube Shorts URLs (youtube.com/shorts/VIDEO_ID) are fully supported and processed identically to regular videos.

How does this compare to the YouTube Transcript API? YouTube does not offer a public transcript API. The official YouTube Data API provides video metadata but not caption text. This scraper extracts the actual transcript content that YouTube displays in its transcript panel — the same text you see when you click "Show transcript" under a video.

What if a video has captions in multiple languages? The scraper returns the transcript in your requested language. If you need transcripts in multiple languages for the same video, run the scraper once per language with the same video IDs and different lang settings.

Can I scrape live stream transcripts? Live streams that have been archived and processed by YouTube (with auto-generated captions) can be scraped like any other video. Active live streams cannot be scraped.

Why do some music videos return no_captions? Many music videos — especially official music videos from major labels — do not have caption tracks enabled on YouTube. The video contains music but YouTube's auto-caption system either doesn't generate captions for music-only content or the uploader has disabled them.

How accurate are auto-generated captions? YouTube's auto-generated captions are typically 85-95% accurate for clear English speech. Accuracy drops for heavy accents, technical terminology, multiple speakers talking simultaneously, and background noise. For research requiring high accuracy, look for videos with manually uploaded captions.

Can I filter results to only include videos that have transcripts? Yes. After the run completes, filter the dataset by status: ok to get only videos with successfully extracted transcripts. Videos without captions are still included in the output (with status: no_captions) so you have a complete audit trail of every video processed.

What is the maximum number of videos I can process? There is no technical limit. The scraper has been tested with 10,000+ videos in a single run. For very large runs, the Apify platform handles memory management and container migration automatically.

Does this scraper support playlists? Not directly. Extract video IDs from a playlist using the YouTube Channel Scraper or a playlist extractor first, then feed those IDs into this scraper.