YouTube Search Results Scraper

Search YouTube by keyword and export ranked public video results for content discovery, competitor research, and recurring topic monitoring. The Actor returns video IDs, titles, channels, URLs, thumbnails, duration, publish timing, visible views, snippets, the originating query, and result position without requiring a YouTube API key or account.

Use one query for a focused snapshot or submit several related queries in one run. Public YouTube filters let you sort by relevance, upload date, view count, or rating and narrow results by upload timing, duration, country, and language.

What does this YouTube search scraper do?

The Actor reproduces public YouTube video search as structured dataset rows. It follows continuation pages until it reaches your limit or YouTube runs out of results. Every row keeps both its position within the query and its position across the full run, making repeated exports easy to compare.

It returns video results only. Channels, playlists, movies, private videos, account-specific history, and personalized signed-in recommendations are outside its scope.

Who is it for?

• Content strategists finding topics, formats, and channels appearing for target keywords.
• Creator teams tracking which videos rank for recurring content themes.
• SEO researchers comparing public search positions, titles, snippets, and visible engagement.
• Competitive intelligence teams collecting consistent snapshots for change analysis.
• Data teams feeding YouTube search results into spreadsheets, warehouses, dashboards, or automation.
• Developers and AI agents that need normalized YouTube discovery data through an API or MCP.

Why use this Actor?

• No Google API project, YouTube Data API key, cookies, or account is required.
• Multiple queries can be processed in one run with separate per-query ranks.
• Supported public search filters can be combined instead of encoded manually.
• Continuation pagination supports more than the first visible page.
• Rows include raw display text and parsed numeric values for views and duration.
• A lightweight HTTP implementation avoids browser startup and media downloads.
• Results are ready for JSON, CSV, Excel, API, webhook, and dataset integrations.

What data can I extract?

Field	Meaning
query	Search term that produced the result
position	One-based position across the complete run
queryPosition	One-based position within that query
videoId	Stable YouTube video ID
title	Public search-result title
url	Canonical YouTube watch URL
channelName, channelId, channelUrl	Channel identity when exposed
thumbnailUrl	Highest-resolution search-result thumbnail exposed
durationText, durationSeconds	Display duration and parsed seconds
publishedTimeText	Relative timing such as 2 weeks ago
viewCountText, viewCount	Visible count text and parsed number
snippet	Description snippet when YouTube includes one
badges	Visible labels such as New, 4K, or Verified
sortBy, uploadDate, duration	Filters applied to the row
region, language	Localization used for the query
scrapedAt	Collection timestamp in ISO 8601 format

Fields that YouTube does not expose for a result are returned as null or an empty badge array. Visible counts are snapshots, not historical metrics.

How to search YouTube and export results

1. Open the Actor input page.
2. Add one or more terms to Search queries.
3. Choose the total result limit and, if needed, a per-query limit.
4. Select a sort order, upload date, and duration filter.
5. Set the two-letter region and language used for public search.
6. Click Start.
7. Open the Dataset tab to preview or export the rows.

A small first run with 20 results is a good way to verify that the selected filters match your research question.

Input parameters

Input	Type	Default	Description
queries	string array	required	One to 20 non-empty search terms
maxItems	integer	100	Total saved rows across all queries, from 1 to 1,000
maxResultsPerQuery	integer	100	Maximum rows for each query, from 1 to 1,000
sortBy	string	relevance	relevance, uploadDate, viewCount, or rating
uploadDate	string	any	any, hour, today, week, month, or year
duration	string	any	any, short, medium, or long
region	string	US	Two-letter country code such as US, GB, DE, IN, or BR
language	string	en	Language code such as en, de, es, or pt-br

YouTube defines short videos as under four minutes, medium videos as four to 20 minutes, and long videos as over 20 minutes.

Example input

{
  "queries": [
    "creator economy news",
    "YouTube SEO 2026",
    "creator brand sponsorships"
  ],
  "maxItems": 60,
  "maxResultsPerQuery": 20,
  "sortBy": "uploadDate",
  "uploadDate": "month",
  "duration": "any",
  "region": "US",
  "language": "en"
}

This input creates three ranked groups in one dataset. Schedule it weekly and compare videoId, queryPosition, and scrapedAt in your own database to identify new entries or movement.

Example output

{
  "query": "AI coding tutorial",
  "position": 1,
  "queryPosition": 1,
  "videoId": "2FJlhoDYNPE",
  "title": "AI Coding Masterclass: From Beginner to Expert in 90 Minutes",
  "url": "https://www.youtube.com/watch?v=2FJlhoDYNPE",
  "channelName": "Riley Brown",
  "channelId": "UCMcoud_ZW7cfxeIugBflSBw",
  "channelUrl": "https://www.youtube.com/@rileybrownai",
  "thumbnailUrl": "https://i.ytimg.com/vi/2FJlhoDYNPE/hq720.jpg",
  "durationText": "1:36:02",
  "durationSeconds": 5762,
  "publishedTimeText": "11 months ago",
  "viewCountText": "239,411 views",
  "viewCount": 239411,
  "snippet": "Learn practical AI coding workflows from setup to deployment.",
  "badges": [],
  "sortBy": "relevance",
  "uploadDate": "any",
  "duration": "any",
  "region": "US",
  "language": "en",
  "scrapedAt": "2026-08-29T20:38:24.869Z"
}

The sample reflects the current output shape. Search rankings and public metadata can change between runs.

How much does it cost to search YouTube videos?

The Actor uses pay-per-event pricing:

• a one-time Run started event of $0.005 per run;
• one Item extracted event for each accepted video row.

The BRONZE item price is $0.001832 per result, with lower per-result rates on higher tiers. At that rate, a 20-result run is approximately $0.04164 including the start event, 100 results are approximately $0.1882, and 1,000 results are approximately $1.837. Your active Apify pricing tier determines the exact item rate shown before the run.

Failed, duplicate, and rejected renderers are not charged as items. Compute is included under the Actor's pay-per-event model where the platform indicates it.

Recurring monitoring workflow

The Actor returns current snapshots rather than maintaining history inside the run. For repeat monitoring:

1. Create a saved Task with stable queries and filters.
2. Schedule it daily, weekly, or monthly in Apify Console.
3. Send completed-run data to a webhook, Google Sheets, a database, or cloud storage.
4. Compare rows by query plus videoId.
5. Use queryPosition to calculate rank movement and scrapedAt to identify the observation.

Keeping history outside the Actor avoids mixing old and new rows in a run's default dataset.

Export and integration options

From the dataset you can download JSON, JSONL, CSV, XML, RSS, or Excel. Common workflows include:

• enriching an editorial calendar with currently ranked titles and channels;
• loading search snapshots into BigQuery, Snowflake, PostgreSQL, or a BI tool;
• triggering a webhook when a scheduled run completes;
• comparing competitor visibility by query and position;
• passing selected video URLs to a separate metadata, comments, or transcript workflow.

Apify integrations also connect runs to Zapier, Make, Slack, Google Drive, GitHub, and other services.

Run through the API

Replace YOUR_TOKEN with an Apify API token.

cURL

curl -X POST \
  "https://api.apify.com/v2/acts/automation-lab~youtube-search-results/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"queries":["OpenAI DevDay"],"maxItems":25,"sortBy":"uploadDate","uploadDate":"year"}'

JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('automation-lab/youtube-search-results').call({
  queries: ['OpenAI DevDay'],
  maxItems: 25,
  sortBy: 'uploadDate',
  uploadDate: 'year',
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

Python

from apify_client import ApifyClient

client = ApifyClient(token="YOUR_TOKEN")
run = client.actor("automation-lab/youtube-search-results").call(run_input={
    "queries": ["OpenAI DevDay"],
    "maxItems": 25,
    "sortBy": "uploadDate",
    "uploadDate": "year",
})
items = client.dataset(run["defaultDatasetId"]).list_items().items
print(items)

Use with MCP and AI agents

Add the Actor to Claude Code:

claude mcp add --transport http apify \
  "https://mcp.apify.com?tools=automation-lab/youtube-search-results"

For Claude Desktop, add this server to the mcpServers object in its configuration. In Cursor, add it under Settings → Tools & MCP. In VS Code, add the same HTTP server through your MCP extension or workspace configuration:

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com?tools=automation-lab/youtube-search-results"
    }
  }
}

Example prompts:

• “Find the top 30 YouTube videos for AI agent tutorial in the US and summarize recurring channel names.”
• “Search these three creator economy topics, sort by upload date, and return videos from the last month.”
• “Compare query positions from this dataset with last week's export.”

Limits and result behavior

• Results reflect public, logged-out YouTube search and may differ from a signed-in person's personalized page.
• YouTube can localize or reorder results over time even with identical inputs.
• Relative publish timing and compact visible views are whatever YouTube exposes in the search result.
• Some results omit snippets, channel URLs, duration, or parsed view counts; those fields are nullable.
• Live streams and upcoming videos may not expose a normal duration or numeric visible view count.
• The Actor caps a run at 20 queries and 1,000 rows.
• A valid query with no matching public videos succeeds with zero rows.
• Network failures, rate limits, or incompatible response changes fail the run rather than silently returning a false empty result.

Reliability and troubleshooting

Requests use a lightweight logged-out YouTube data route with bounded retries for temporary network errors, HTTP 429, and server errors. Stable invalid requests are not retried.

Why did I get fewer rows than requested? YouTube may have fewer video results matching the combined filters, may stop offering a continuation, or may omit unsupported result cards. Try a broader query or less restrictive upload-date and duration filters.

Why are the rankings different from my browser? Signed-in history, cookies, experiments, location, and personalization affect the consumer website. This Actor uses the supplied public region and language in a logged-out context for repeatable collection.

Why is viewCount null? Upcoming streams and some result formats expose text that is not a numeric view total. Check viewCountText for the original visible value.

Legality and responsible use

This Actor collects publicly visible search-result metadata. Use it for legitimate research and automation, respect YouTube's terms and applicable law, and avoid collecting or combining data in ways that infringe privacy, intellectual property, or contractual rights. You are responsible for your inputs, schedules, retention, and downstream use.

The Actor does not bypass authentication, access private videos, download media, or retrieve account-specific data.

FAQ

Does it need a YouTube Data API key? No. It uses public logged-out YouTube search data.

Can it search several keywords? Yes. Supply up to 20 queries and control both the total and per-query limits.

Can I filter to videos uploaded this week? Yes. Set uploadDate to week.

Can I export Shorts only? Not currently. The Actor returns YouTube video search results and supports public duration filters, but it does not claim a Shorts-only mode.

Does it preserve rank history? Each row contains current positions and a timestamp. Store scheduled-run datasets in your own destination to calculate historical changes.

Are visible views exact? When YouTube exposes an exact count it is parsed. Compact counts are normalized approximately, and the original text remains in viewCountText.

Related YouTube Actors

Use other automation-lab Actors when your workflow starts from a known entity rather than a search term:

• YouTube Channel Videos Scraper for a channel's public video list.
• YouTube Channel Search Scraper for searching within a known channel.
• YouTube Comments Scraper for public comments on selected videos.
• YouTube Transcript Scraper for transcript extraction after discovery.
• YouTube Thumbnail Downloader for storing thumbnail assets from known video URLs.