YouTube Shorts and Search Data for Content Research
Find what ranks on YouTube for a topic, which Shorts get views and engagement, and which creators keep showing up, using search results and per Short stats without an API key.
The actor referenced in this article. Pay only for results delivered.
Before you script a video, you want three answers: what already ranks for the topic, which short videos on it are pulling views, and who keeps making them. YouTube Studio only shows your own channel. The search page rounds view counts, hides likes, and no longer sorts by date. Opening fifty Shorts one by one to note their numbers takes an afternoon and goes stale in a week. Two actors cover this: one reads YouTube search the way a visitor sees it, the other looks up each Short for its exact numbers.
Try it live: YouTube Shorts Scraper. Pay per result delivered.
Which actor answers which question
| Question | Actor | Input | Exact likes and comments |
|---|---|---|---|
| Which Shorts on a topic or channel get views and engagement | YouTube Shorts Scraper | Channels, keywords or Short links | Yes |
| What ranks for a keyword: videos, Shorts, channels, playlists | YouTube Search Scraper | Keywords or search links | No |
| How each Short does against its channel’s median | YouTube Shorts Views Scraper | Channels | Views with a reach ratio |
| What viewers actually say | YouTube Comments Scraper | Videos or Shorts | Comment text |
| Business emails of the creators you found | YouTube Email Finder | Channels | Not applicable |
Most research starts with the first two. Search tells you what YouTube puts in front of people; the Shorts actor tells you how those people reacted.
What gets views in a niche this month
{
"searchQueries": ["home workout", "meal prep"],
"searchSort": "popularity",
"uploadDate": "month",
"minViews": 10000,
"maxShortsPerInput": 200
}
This is the YouTube Shorts Scraper. Each keyword returns Shorts only, most viewed first, uploaded this month, with at least 10,000 views. minViews is checked against the view count on the listing before the per Short lookup, so Shorts that cannot reach it are never looked up or charged. YouTube serves up to about 400 Shorts per keyword: 409 in the actor’s test.
Every row carries views, likes and comment_count as exact numbers, engagement_rate worked out as likes plus comments over views, published_at to the second, duration_seconds, hashtags, tags, category, thumbnail_url, and the channel: channel_name, channel_handle, channel_url and channel_subscribers. Group by channel_handle for a creator list, or count hashtags to see what the niche is tagging this month.
To follow specific competitors instead, give channels and a date cutoff:
{
"channels": ["@competitor_one", "@competitor_two", "https://www.youtube.com/@competitor_three"],
"oldestPostDate": "7 days",
"maxShortsPerInput": 100
}
On a channel, reading stops at the first Short older than the cutoff. A scheduled channel with nothing new costs 3 small requests and no charge. channelSort: "popular" gives a creator’s best Shorts instead, but it is ignored while a date cutoff is set.
What ranks for a keyword
{
"searchQueries": ["iphone 17 review"],
"uploadDate": "week",
"sortBy": "date",
"maxResultsPerKeyword": 50
}
This is the YouTube Search Scraper. YouTube removed its upload date sort in 2025, so sortBy: "date" makes the actor read every page YouTube serves for the keyword (up to 25) and return the newest first. In its test this search read 25 pages, 393 videos from that week, and returned the 50 newest.
Other settings that matter for research:
typepicksvideo,shorts,channel,playlistorall. Withchannel, each row carrieschannel_handle,subscribersandchannel_is_verified: a creator list for a topic in one run.countryandlanguageset the market YouTube ranks for. In the actor’s tests only 8 to 12 of the top 20 stayed the same between English and the local language, so check each market you publish in.deduplicate: falsereturns a video once per keyword with itspositionfor each, which is what rank tracking needs. Comparepositionbyvideo_idbetween scheduled runs.excludeShorts: trueleaves out Shorts YouTube mixes into video results, and they are not charged.
Search rows carry position, keyword, title, url, views, published_text as shown (“2 months ago”) with published_at_estimate and published_precision, duration_seconds, description_snippet, chapters, badges and flags such as is_short, is_4k and has_captions. They do not carry likes or comment counts; the search page does not show them.
Python: find the creators for a topic, then their best Shorts
This script uses search to find channels for a topic, then pulls each channel’s most popular Shorts and summarises what works.
pip install apify-client
from collections import Counter, defaultdict
from statistics import median
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
# 1. Channels YouTube shows for the topic
search_run = client.actor("themineworks/youtube-search-scraper").call(run_input={
"searchQueries": ["meal prep", "high protein recipes"],
"type": "channel",
"maxResultsPerKeyword": 40,
"country": "US",
"language": "en",
})
handles = []
for row in client.dataset(search_run["defaultDatasetId"]).iterate_items():
if "_type" in row or row.get("result_type") != "channel":
continue
if row.get("channel_handle"):
handles.append(row["channel_handle"])
print(f"{len(handles)} channels found")
# 2. Each channel's most popular Shorts
shorts_run = client.actor("themineworks/youtube-shorts-scraper").call(run_input={
"channels": handles[:50],
"channelSort": "popular",
"maxShortsPerInput": 20,
})
by_channel = defaultdict(list)
tags = Counter()
for short in client.dataset(shorts_run["defaultDatasetId"]).iterate_items():
if "_type" in short: # summary and error rows
continue
by_channel[short["channel_handle"]].append(short)
tags.update(short.get("hashtags", []))
report = []
for handle, rows in by_channel.items():
rates = [r["engagement_rate"] for r in rows if "engagement_rate" in r]
report.append({
"channel": handle,
"subscribers": rows[0].get("channel_subscribers"),
"median_views": median(r["views"] for r in rows),
"median_engagement": round(median(rates), 4) if rates else None,
"median_length_s": median(r["duration_seconds"] for r in rows),
"top_short": max(rows, key=lambda r: r["views"])["url"],
})
for r in sorted(report, key=lambda r: r["median_views"], reverse=True)[:15]:
print(r)
print("Most used hashtags:", tags.most_common(20))
The search actor delivers each channel once even when both keywords find it, because deduplicate is on by default. On the Shorts side, views, duration_seconds and hashtags were present on every row of a 1,211 row test, so the script can rely on them. likes is left out when a channel hides likes, which is why engagement_rate is read with a check.
What it costs
YouTube Shorts Scraper: $0.001 per Short on the Apify Free plan, $0.0008 on Bronze, $0.0007 on Silver and $0.0006 on Gold and above. A $0.001 start fee is charged once per run, after the first Short row is stored, so a run that delivers nothing costs nothing. Channels that do not exist, links that are not Shorts, Shorts found twice and Shorts under minViews or older than oldestPostDate are free.
YouTube Search Scraper: $0.0009 per result on the Free plan, $0.00054 on Bronze, $0.00045 on Silver and $0.00036 on Gold and above, the same for videos, Shorts, channels and playlists. A $0.001 start fee is charged once per run, after the first result row is stored. Duplicates while deduplicate is on, rows outside your upload window, Shorts dropped by excludeShorts and the extra pages read for newest first are free.
For the script above on the Gold tier: at most 80 channel rows cost up to 80 × $0.00036 = $0.0288, and 50 channels × 20 Shorts = 1,000 Shorts cost $0.60, plus two start fees of $0.001.
Limits worth knowing
- Search has no engagement. Likes and comment counts come only from the Shorts actor’s per Short lookup.
- Upload dates in search are estimates. The page shows an age;
published_precisionsays whether it is accurate to the hour or only to the year. The Shorts actor givespublished_atto the second. - Subscribers are rounded in both actors, because YouTube shows that figure rounded to everyone.
- YouTube caps result depth. Video searches ended between 330 and 390 results in tests, Shorts searches on the search page near 180, and the Shorts actor’s keyword search at about 400.
- Partial rows are possible. YouTube holds back part of its answer on many server IPs. The Shorts actor moves to IPs that answer in full, and 1,211 of 1,211 rows were complete in its largest test, but a row marked
data_status: "partial"has no length or exact time. - Metadata only. Neither actor downloads video or audio.
Related
Explore the scraper referenced in this article: inputs, outputs, and pricing, then run it on Apify.
Frequently asked questions
Which actor gives likes and comment counts? +
YouTube Shorts Scraper. It looks up every Short on its own, so views, likes and comment count are exact whole numbers. YouTube Search Scraper returns what the search page shows: views, upload age, length and badges, but no likes or comment counts.
Do I need a YouTube Data API key or quota? +
No. Both actors read YouTube as a logged-out visitor sees it, with no account, cookies, API key or quota. Apify's datacenter proxy is included in the price.
How many results can I get for one keyword? +
YouTube stops serving a search after a while. In tests a video search ended between 330 and 390 results and a Shorts search on the search page near 180. The Shorts actor's own keyword search reached 409 Shorts. Add related keywords to go further.
Can I sort search results by upload date? +
YouTube removed that sort in 2025. YouTube Search Scraper rebuilds it: with sortBy set to date it reads every page YouTube serves for the keyword, up to 25, and returns the newest results first by the upload age shown on each.
Find Instagram Accounts by Keyword, Then Get the Emails and Phones They Publish
Search public Instagram posts and reels by keyword to find the accounts active in a niche, then collect the emails, phones and WhatsApp numbers those accounts publish in their bio, link in bio page and website.
Export Every Image and Caption From Instagram Carousel Posts
Export every slide image link, slide type and caption from Instagram carousel posts to a CSV, by profile or by post link, with Instagram Post Scraper.
Export a YouTube Video's Comments and Replies to CSV
Pull a YouTube video's comments and their replies into one CSV with YouTube Comments Scraper, each thread rebuilt from parent_comment_id.
Monitor the Newest Comments on Your YouTube Videos Every Day
Read the newest comments on your YouTube videos every day with YouTube Comments Scraper, keep only unseen ones, and flag questions you have not answered.