YouTube Research & Transcript Scraper

Search YouTube, export channel and video data, and pull transcripts for shortlisted videos without managing the YouTube Data API.

This actor is built for a shortlist-first workflow:

• discover videos by keyword
• snapshot one or more channels
• enrich a shortlisted set with publish date, category, and fuller metadata
• pull transcripts only for the videos you actually need

That makes it a better fit for research, content analysis, LLM ingestion, and competitor monitoring than a one-price-fits-all commodity scraper.

What you can do with it

1. Build a research shortlist from search

Use searchQueries when you want to map a topic quickly and export a clean table of candidate videos.

{
  "searchQueries": ["ai workflow automation"],
  "maxResults": 50
}

Typical output fields:

• videoId
• title
• url
• channelName
• channelId
• viewCount
• duration
• publishedText
• description
• thumbnailUrl
• source

This is the lightest tier and the best first pass when you are still exploring.

2. Snapshot one or more channels

Use channelUrls to export the latest videos from a creator, company, competitor, or media outlet.

{
  "channelUrls": ["https://www.youtube.com/@freecodecamp"],
  "maxResults": 100
}

This is useful for:

• competitor channel monitoring
• editorial audits
• recurring content snapshots
• identifying which videos deserve deeper enrichment

3. Enrich a shortlisted set with stronger metadata

Turn on scrapeDetails when you need a more research-ready record for each item.

{
  "searchQueries": ["legal ai"],
  "maxResults": 20,
  "scrapeDetails": true
}

Detail enrichment is the middle tier. It fills in fields such as:

• publishedAt
• category
• richer description
• refreshed viewCount

4. Pull transcripts for selected videos

Use videoUrls plus includeTranscript when you already know which videos matter and want timestamps and full text.

{
  "videoUrls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],
  "includeTranscript": true,
  "transcriptLanguage": "en"
}

Transcript rows can include:

• transcript
• transcriptLanguage
• transcriptText
• enrichmentLevel

The actor only adds the transcript tier when transcript segments are actually returned.

Input fields

You can run the actor with any combination of these inputs:

Field	Type	What it does
searchQueries	array	Search YouTube by keyword for discovery and shortlist building
channelUrls	array	Export videos from one or more channels
videoUrls	array	Target specific videos directly
maxResults	integer	Caps search and channel result volume
scrapeDetails	boolean	Adds richer metadata for each row
includeTranscript	boolean	Attempts transcript extraction
transcriptLanguage	string	Preferred transcript language, with fallback if unavailable

Output contract

Every row includes the core video record, plus two fields that make downstream handling easier:

• recordVersion — current output contract version
• enrichmentLevel — base, detail, or transcript

That means downstream automations can branch cleanly without guessing which tier produced the row.

Pricing

This actor uses tiered pay-per-event pricing.

Event	Price
Actor start	$0.00005
Base video row	$0.0015
Detailed video row	$0.0025
Transcript-ready row	$0.0040

Two practical notes matter:

• transcript rows are priced separately because transcript extraction is materially heavier than plain metadata collection
• this actor is configured so Apify bills platform usage directly to the user, which keeps creator-side economics clean but means heavy transcript runs cost more than lightweight discovery runs

The intended workflow is:

1. run search or channel discovery cheaply
2. shortlist what you actually need
3. run detail or transcript mode only on that shortlist

That is the fastest way to keep both runtime and spend under control.

What makes this different

This actor is not trying to hide the difference between cheap and expensive work.

Search and channel listing are lightweight. Metadata enrichment is heavier. Transcript extraction is the heaviest path.

Instead of pretending those modes should all be sold at one flat price, this actor exposes the tiers directly so the economics match the actual work being done.

Common use cases

• competitor content tracking
• creator and channel research
• sourcing YouTube examples for sales, research, and consulting
• transcript collection for LLM or RAG pipelines
• building structured datasets for media analysis
• shortlist-first workflows where broad discovery and deep extraction should not be priced the same way

What this actor does not promise

• transcripts for every video
• a single flat cost across discovery and transcript-heavy runs
• full parity with every endpoint in the official YouTube API

It is best when you want fast research-oriented extraction without managing API quota, and when you are willing to run heavy transcript jobs selectively instead of blindly across large batches.

Example workflow

Start broad:

{
  "searchQueries": ["warehouse automation ai"],
  "maxResults": 30
}

Then enrich the videos you actually care about:

{
  "videoUrls": [
    "https://www.youtube.com/watch?v=XVv6mJpFOb0",
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  ],
  "scrapeDetails": true,
  "includeTranscript": true,
  "transcriptLanguage": "en"
}

That split workflow is what this product is optimized for.