YouTube Outlier Finder

Find breakout YouTube videos on small channels — videos massively outperforming their own channel's baseline, the earliest public signal that the algorithm is pushing a topic.

An outlier is a video on a small channel doing 5x, 20x, even 100x the channel's typical views. That gap isn't the channel's own audience showing up — it's the YouTube algorithm testing a topic on cold viewers it has never served the channel before. Spot those breakouts and you see which formats and niches are heating up before they get crowded.

Most tools dump 1,000 raw search rows and leave the analysis to you. This actor does the opposite: it applies a three-gate filter, scores every survivor, and returns a ranked shortlist of confirmed breakouts — insight, not a data dump. No API key, no Google Cloud project, no quota. SaaS research tools like VidIQ and 1of10 sell this exact signal for $30–50/month; here you pay per scan.

You give it	You get back
Search queries (topics/niches)	A ranked list of breakout videos
An upload window (week / month / year)	Each scored: views ÷ channel baseline
View, score, and subscriber gates	Only videos that clear every gate
A results cap	Sorted by outlier score, highest first

What does the YouTube Outlier Finder do?

It runs each of your search queries against YouTube's own Innertube API, collects fresh videos in your chosen upload window, then does the analysis a human researcher would do by hand:

1. Collects candidates across every query and deduplicates them, keeping only videos above your absolute view floor.
2. Fetches each unique channel once to read its subscriber count and the view counts of its recent uploads.
3. Computes a baseline — the median views of the channel's recent uploads, excluding the candidate video itself so its own virality can't inflate the number it's judged against.
4. Gates and scores — drops channels above your subscriber cap, computes outlierScore = views ÷ baseline, and discards anything below your minimum score.
5. Ranks and returns the top breakouts, sorted by score descending.

The result is a confirmed shortlist: every row is a small-channel video that demonstrably beat its own track record by the multiple you asked for.

Who is it for?

Audience	What they pull from it
YouTubers & creators	Proven formats and hooks the algorithm is rewarding in their niche right now
Content agencies	Weekly breakout reports across a roster of client niches, ready for a deck
Niche researchers	Evidence that small channels in a niche can break out at all — before betting on it
Trend scouts	An early map of where attention is moving across several topics at once
AI content pipelines	Clean, ranked rows to feed an agent that drafts ideas, titles, or scripts from winners

Why use this tool?

• Insight, not raw dumps. Search scrapers hand you thousands of rows to sift. This actor returns a ranked shortlist of breakouts with the scoring already done.
• Three-gate filter, minimal noise. Absolute view floor + outlier multiple + channel-size cap. A video has to clear all three to show up.
• No API key, no quota. Runs on YouTube's public Innertube API — no Google Cloud setup, no YouTube Data API key, no daily quota ceiling.
• Baseline that can't be gamed. The channel baseline excludes the candidate video and uses the median (not the average), so one past viral hit can't hide today's breakout.
• Built by a practitioner. The gates and defaults come from running this exact detection pipeline for real channel research, not from guesswork.

What data do you get?

Each dataset row is one confirmed outlier:

{
  "outlierScore": 6.4,
  "videoId": "dQw4w9WgXcQ",
  "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "title": "The Hidden Cost of Letting AI Write Your Code",
  "views": 14064,
  "published": "4 days ago",
  "length": "11:48",
  "channelName": "Uma Abu",
  "channelUrl": "https://www.youtube.com/channel/UC_x5XG1OV2P6uZZ5FSM9Ttw",
  "channelSubs": 26000,
  "channelBaselineViews": 2200,
  "channelRecentUploads": 15,
  "matchedQuery": "ai tools"
}

Field	Meaning
outlierScore	Views ÷ channel baseline. 6.4 means the video did 6.4x what this channel typically does
videoId	YouTube video ID
url	Direct watch URL
title	Video title
views	View count at scan time
published	Relative upload time, e.g. 4 days ago
length	Video duration, e.g. 11:48
channelName	Channel name
channelUrl	Direct channel URL
channelSubs	Channel subscriber count (null if hidden)
channelBaselineViews	Median views of the channel's recent uploads — the number the score is measured against
channelRecentUploads	How many recent uploads went into the baseline
matchedQuery	Which of your search queries surfaced this video

How to find viral YouTube videos on small channels

Set maxSubscribers to keep only small channels (default 100,000), pick a fresh uploadWindow like week or month, and raise minOutlierScore to demand a bigger over-performance multiple. The actor filters out big-brand channels — where views come from an existing audience, not the topic — and surfaces small channels whose latest upload suddenly outran everything they've posted. That spike is the algorithm distributing the video beyond the channel's subscribers, which is exactly the "viral on a small channel" pattern.

How to find YouTube content ideas

Point queries at the topics you cover and read the returned titles, formats, and lengths. Because every row already beat its own channel's baseline by your chosen multiple, you're looking at hooks and angles the algorithm is actively rewarding — not just popular videos from channels that are popular anyway. Sort by outlierScore, open the top results, and reverse-engineer what made them break out. Feed the rows to an AI agent (see the MCP section) to generate title and script variations from proven winners.

How to discover trending YouTube topics early

Run several queries across adjacent niches with uploadWindow: "week". Fresh windows surface videos while they're still climbing, before the trend is obvious enough to be crowded. A cluster of high-score outliers around a theme across different small channels is an early signal that attention is shifting there. Re-run the same query set on a schedule and watch which themes keep producing breakouts week over week.

How much does it cost

There's no YouTube API key and no per-request quota to buy. You pay only for the Apify platform compute of each run, and the actor is designed to be economical: it fetches each unique channel once (not per video), applies gates before scoring, and returns a small ranked shortlist rather than raw search pages. Compared with VidIQ, 1of10, and similar outlier-research subscriptions at $30–50/month, you run it on demand and only when you need a fresh scan.

How to validate a YouTube niche before starting a channel

Before committing to a niche, run the actor with queries describing it and a low maxSubscribers cap. If small channels in that niche are producing outliers, newcomers can break through on topic strength alone — the algorithm is willing to push unknown channels there. If you get almost nothing back even with a wide uploadWindow and a modest minOutlierScore, that's a signal the niche is saturated by established brands and hard to enter cold. It's cheap evidence for a decision that otherwise takes months to learn the hard way.

How to export YouTube outliers to CSV or Google Sheets

Every run writes to an Apify dataset. From the run's Storage tab you can export the results to CSV, Excel, JSON, or XML in one click, or pull them programmatically via the Apify API and Dataset endpoints. To land rows in Google Sheets, connect the dataset through Apify integrations (Make, Zapier) or fetch the dataset JSON and import it. The dataset also ships with a clean Overview table view — score, title, views, channel, subs, baseline, published, length, and URL — for quick scanning without exporting.

YouTube trend discovery API alternative

If you were reaching for the YouTube Data API to build trend discovery, this actor is a lighter path: no Google Cloud project, no OAuth, no API key, and no daily quota that caps how much you can scan. It talks to YouTube's public Innertube endpoints directly and hands back structured, scored rows. Call it as an API or as an MCP tool, wire it into an agent, and get outlier detection without managing API credentials or rationing quota.

Is it legal to scrape YouTube?

This actor reads only publicly visible data — the same search results and channel pages any signed-out visitor sees — and does not log in, bypass access controls, or touch private information. Public data collection is broadly permissible, but you are responsible for your own use: review YouTube's Terms of Service and your local laws before running at scale, and don't republish content in ways that infringe creators' rights. See the disclaimer below.

Input configuration

Field	Type	Default	Description
queries	string[]	(required)	Topics or niches to scan, e.g. ai tools, homestead off grid, woodworking asmr. Each is searched separately, then results are deduplicated
uploadWindow	string	month	Only consider videos uploaded within this window: week, month, or year. Fresher windows surface earlier signals
duration	string	any	any includes everything; long restricts to 20+ minute videos — the strongest signal for serious niches
minViews	integer	5000	Absolute view floor. A video needs at least this many views regardless of channel size (min 100)
minOutlierScore	integer	5	Minimum multiple over the channel baseline. 5 = the video did 5x the channel's typical views (min 2, max 100)
maxSubscribers	integer	100000	Ignore channels above this size — small channels are the purest topic signal (min 1,000)
maxResults	integer	50	Cap on outliers returned, sorted by score descending (min 1, max 500)

FAQ

How is the baseline computed? It's the median views of the channel's ~15 most recent uploads, with the candidate video's own views removed from the pool first, and floored at 1,000 so a near-empty channel can't produce a junk score. The median (not the average) means one previous viral hit doesn't drag the baseline up and mask a new breakout. A channel needs at least two other recent uploads to score; otherwise it's skipped.

Why did I get only a few results? Outliers are rare by design — that's the entire point. Every row cleared three gates. To widen the net, extend uploadWindow, add more queries, lower minOutlierScore, or drop minViews.

Do I need a YouTube account or API key? No. No login, no API key, no OAuth, and no quota. It runs on YouTube's public Innertube API.

Integrations

Connect the Actor to the tools you already use through Apify integrations:

• Outliers → Google Sheets — auto-append each scan's breakout videos into a running content-idea board.
• Outliers → Slack / Discord — get notified when a fresh breakout appears in your niche.
• Outliers → Make / Zapier — feed winning videos into a content pipeline, a Notion board, or an AI script generator.
• Outliers → BigQuery / Snowflake — warehouse outlier data via Airbyte to track topic momentum over time.
• Scheduled runs & webhooks — run a weekly niche scan automatically, or POST results the moment a scan completes.

Use with AI agents via MCP

The actor is available as a tool over the Model Context Protocol, so agents can call it directly:

claude mcp add --transport http apify "https://mcp.apify.com?tools=datascoutlab/youtube-outlier-finder"

Example prompt: "Find breakout videos on small channels for 'ai tools' and 'local llm' from the last week and summarize the common hooks."

Related scrapers

• App Store Reviews Scraper — reviews from 175 countries with filters and clean JSON
• Google Play Reviews Scraper — Android app reviews from any country, no API key
• Google Trends Scraper — interest-over-time data that actually works
• Telegram Channel Scraper — messages, views, and keyword search across channels

Disclaimer

This actor is not affiliated with, endorsed by, or sponsored by YouTube or Google LLC. YouTube is a trademark of Google LLC. It collects only publicly available data and does not access private or restricted content. You are responsible for complying with YouTube's Terms of Service and applicable laws in your jurisdiction.

SEO Keywords

youtube outlier finder, youtube outlier detection, find viral youtube videos small channels, youtube breakout videos, youtube content ideas, discover trending youtube topics, youtube trend discovery api, youtube niche validation, youtube video scraper no api key, youtube innertube api, small channel outliers, youtube algorithm signal, vidiq alternative, 1of10 alternative, youtube outlier score, youtube research tool, early youtube trends, viral video finder, youtube channel baseline, export youtube outliers csv