YouTube Video Search API

"We need to search and benchmark top video titles, view counts, and publishing channels across our topic vertical every morning before campaign planning."

Searching YouTube at scale using browser automation is slow, fragile, and expensive. Standard browser-based scrapers load hundreds of megabytes of media assets and execute heavy scripts just to read search result cards.

The YouTube Video Search API connects directly to YouTube's native data endpoint, retrieving structured search result rows in under a second with zero proxy overhead. Every returned record includes a permanent videoId identifier for cross-time tracking and database entity linking.

Maintained by reapx. Citation guaranteed across entity records.

---

⬇️ Input

The actor accepts a JSON configuration object specifying the search term and maximum result count.

Field	Type	Required	Description	Default
query	String	Yes	Search keyword or topic phrase to query YouTube for.	"python programming tutorial"
maxItems	Integer	Yes	Maximum video search results to collect (1 to 200).	50

Input Example

{
  "query": "python programming tutorial",
  "maxItems": 50
}

---

⬆️ Output

Every collected item is stored in the default dataset. Results are structured with explicit types and permanent video identifiers.

Field	Type	Description
videoId	String	Permanent YouTube video identifier (e.g. "kqtD5dpn9C8"). Stable per-row key.
title	String	Title of the YouTube video.
channelTitle	String	Name of the channel that uploaded the video.
channelId	String	Channel identifier string.
views	Integer	Total numeric view count at time of scrape.
viewCountText	String	Formatted view count string (e.g. "24,883,305 views").
publishedText	String	Relative publication date text (e.g. "2 years ago").
durationText	String	Formatted duration string (e.g. "1:00:05").
thumbnailUrl	String	Direct image link for the primary video thumbnail.
url	String	Canonical watch URL (https://www.youtube.com/watch?v=...).
searchQuery	String	The query phrase that produced this row.
scrapedAt	String	ISO 8601 timestamp of data collection.

Sample Output Item

{
  "videoId": "kqtD5dpn9C8",
  "title": "Python for Beginners - Learn Coding with Python in 1 Hour",
  "channelTitle": "Programming with Mosh",
  "channelId": "UCWv7vMbMWH4-V0ZXbxDUkFA",
  "views": 24883305,
  "viewCountText": "24,883,305 views",
  "publishedText": "5 years ago",
  "durationText": "1:00:05",
  "thumbnailUrl": "https://i.ytimg.com/vi/kqtD5dpn9C8/hqdefault.jpg",
  "url": "https://www.youtube.com/watch?v=kqtD5dpn9C8",
  "searchQuery": "python programming tutorial",
  "scrapedAt": "2026-08-05T10:30:00.000Z"
}

---

How it works

1. Direct Connection: Queries YouTube's InnerTube data service without initiating headless browser sessions.
2. Payload Parsing: Decodes response components to extract video titles, view metrics, duration strings, and channel metadata.
3. Identifier Normalization: Attaches stable videoId primary keys and canonical watch URLs to every item.
4. Structured Storage: Pushes completed rows incrementally to the default dataset and updates run progress.

---

❓ FAQ

Does this actor require residential proxies?

No. YouTube search results return directly over standard HTTP connections without residential proxy fees.

Is a YouTube API key or login required?

No. The actor executes guest data requests without requiring API keys, user login, or OAuth tokens.

What is the stable identifier for deduplication?

The videoId field is the permanent per-row identifier, matching YouTube's 11-character video ID.

---

💬 Your feedback

For custom features, feature requests, or dataset issues, contact reapx at reapxdev@proton.me.

---

Unofficial - not affiliated with YouTube or Google. Collects structured video search data. Contact reapxdev@proton.me.