YouTube Search Scraper - Video Metadata

YouTube Search Scraper: video metadata for your keywords

Turn YouTube keyword searches into a compact dataset of video titles, channel links, displayed views, durations and snippets. Use it to collect source videos for topic research, review a niche's visible search results or feed a video-research workflow.

This Actor reads the initial YouTube search page for each query. It needs no YouTube API key or account credentials. It does not load continuation pages, extract transcripts or promise a complete list of matching videos.

Start with one search

1. Enter a specific query, such as n8n CRM automation tutorial.
2. Set a small result cap and review the run's maximum charge before starting.
3. Open the dataset and check the source videos and view-count text.
4. Check the log if a query returns few or no rows; source blocking and parsing failures are reported there.

{
  "searchQueries": ["n8n CRM automation tutorial"],
  "maxResults": 20,
  "useProxy": true
}

Apify provides JSON, CSV and Excel exports, API access, saved tasks and schedules. Scheduling repeats the search; this Actor does not retain history or identify changes between runs.

Input

Field	Meaning
searchQueries	Required list of search terms. Each nonempty query is requested separately.
maxResults	Maximum video rows per query; default 30, form range 1–200. The initial page usually exposes fewer than a large requested cap.
useProxy	Use Apify's AUTO proxy group; default true. Availability and usage charges depend on the account and current pricing configuration.

The request uses English language headers. There are no input controls for country, publication-date filters, duration filters or authenticated personalization. A query matching several searches can appear more than once because records retain their originating query.

Output fields

Field	Contents
query	Original search query
videoId, url, title	Public video identity and link
channel, channelUrl	Displayed channel name and available channel link
viewCountText	Original displayed view-count label
views	Number parsed from that label; abbreviated counts remain approximate
publishedText	Source publication label, often relative; not an exact timestamp
lengthText	Displayed duration when available
description	Search-result snippet, not the full video description

The Actor collects supported videoRenderer entries. Other search-card layouts, dedicated Shorts shelves, channel cards and playlist cards are not extracted. Missing fields remain null. For example, a source label of 1.2M views is parsed as 1200000; it is not an exact counter read.

Pricing and limits

This Actor charges an Actor-start event per allocated GB of memory, with a minimum of one event. There is no separate per-result event. Open the Pricing tab for the active rate and any subscription-tier discounts, and check memory in the run settings before starting. Startup can be charged even if the run returns no usable results. Output limits cap collection; they do not guarantee coverage.

Troubleshooting

A run can finish after individual queries fail. A green run status does not establish that every query returned useful data. Read warnings for consent pages, unsupported source HTML, HTTP errors or access restrictions; compare the dataset's query values with your input.

Search results vary with time and source behavior. This is a bounded snapshot, not a market-size estimate, historical rank tracker or forecast of viral performance. To report an issue, include the run URL and query in this Actor's Issues tab, without secrets.