Fake Follower Checker: Creator Authenticity Score (YT/IG/TT) avatar

Fake Follower Checker: Creator Authenticity Score (YT/IG/TT)

Pricing

from $15.00 / 1,000 audit completeds

Go to Apify Store
Fake Follower Checker: Creator Authenticity Score (YT/IG/TT)

Fake Follower Checker: Creator Authenticity Score (YT/IG/TT)

Creator handles, your metrics or a scraper dataset in; an explainable 0-100 authenticity score with tier benchmarks out. $0.015 per scored profile.

Pricing

from $15.00 / 1,000 audit completeds

Rating

0.0

(0)

Developer

Creator Fusion

Creator Fusion

Maintained by Community

Actor stats

1

Bookmarked

78

Total users

1

Monthly active users

4 days ago

Last modified

Features

Limited permissionsSingle-tenantServer interface

Share

This Fake Follower Checker gives any YouTube, Instagram or TikTok creator an explainable 0-100 audience authenticity score, from your own metrics, a scraper dataset you already have, or a public YouTube channel, with every signal, its tier benchmark and its weight shown, for $0.015 per scored profile.

Who it's for

  • Brand marketers vetting a shortlist before signing, who want to see views far below subscriber counts.
  • Agencies that already run Instagram/TikTok/YouTube scrapers and want to score 500 profiles from an existing dataset without paying to scrape again.
  • Creators who want a "high authenticity" band for a media kit.
  • AI agents (MCP) that need a single call returning authenticityScore and band.

What's new in v2

  • Explainable score. Five weighted signals (engagement vs tier, views per follower, engagement spikes, comment/like ratio, follower-growth steps). Each item lists value, benchmark, weight, sub-score and contribution. The output speaks of risk signals and a band, never accusations.
  • Bring your own data. Pass profiles[] with followers and recent posts, or the datasetId of a dataset from apify/instagram-profile-scraper, clockworks/tiktok-scraper / tiktok-profile-scraper or streamers/youtube-channel-scraper. No re-scraping.
  • YouTube without keys. A handle is enough: subscriber count from the channel page, latest uploads from the public channel Videos tab, with exact views and likes read from each scored video's watch page (RSS feed as a fallback). Add your own YouTube Data API key (optional) for comment counts.
  • Standby API. GET /score and POST /score for agents and apps (see below).
  • Errors are separate and free. Failed profiles go to the run's separate errors dataset (linked from the Output tab) and are never charged. The default dataset only contains scored profiles.
  • Same price. $0.015 per scored profile ("Audit completed"). The old $0.03 start charge is now the platform default ($0.00005). No new paid events.
  • Instagram / TikTok direct fetching is unofficial, best effort (see Limits). Legacy v1 inputs still work.

Sample output

One scored item from the offline fixtures on the RSS fallback path (real captured @LinusTechTips channel page and RSS feed; on Apify the same fields come from the Videos tab and watch pages with "dataSource": "youtube_page"). You can download the dataset as JSON, CSV, Excel or HTML.

{
"status": "ok",
"platform": "youtube",
"handle": "@LinusTechTips",
"username": "LinusTechTips",
"channelId": "UCXuqSBlHAE6Xw-yeJA0Tunw",
"followers": 16900000,
"followersApproximate": true,
"tier": "mega",
"postsAnalyzed": 12,
"authenticityScore": 93,
"band": "high",
"confidence": "medium",
"dataSource": "youtube_rss",
"scoringVersion": "1.0",
"unofficial": false,
"signals": [
{"key": "er_vs_tier", "value": 0.015, "benchmark": [0.01, 0.04], "benchmarkSource": "provisional", "weight": 0.3, "subScore": 100, "contribution": 40},
{"key": "views_per_follower", "value": 0.1, "benchmark": [0.015, 0.25], "benchmarkSource": "provisional", "weight": 0.25, "subScore": 100, "contribution": 33.3},
{"key": "engagement_spikes", "value": 1, "benchmark": [0, 1], "benchmarkSource": "rule", "weight": 0.2, "subScore": 75, "contribution": 20},
{"key": "comment_like_ratio", "value": null, "benchmark": null, "benchmarkSource": null, "weight": 0, "subScore": null, "contribution": null, "note": "comments unavailable"},
{"key": "follower_growth_steps", "value": null, "benchmark": null, "benchmarkSource": null, "weight": 0, "subScore": null, "contribution": null, "note": "no follower history supplied"}
],
"engagement": {"medianViews": 1693253, "medianLikes": 22730, "medianComments": null, "medianEngagementRate": 0.0148, "postsPerWeek": 4.81},
"redFlags": ["1 post with likes or comments > 3× median"],
"redFlagsText": "1 post with likes or comments > 3× median",
"topRedFlag": "1 post with likes or comments > 3× median",
"warnings": [],
"sig_er_value": 0.015,
"sig_er_subScore": 100,
"sig_vpf_value": 0.1,
"sig_vpf_subScore": 100,
"sig_spikes_value": 1,
"sig_spikes_subScore": 75,
"sig_clr_value": null,
"sig_clr_subScore": null,
"sig_growth_value": null,
"sig_growth_subScore": null
}

Benchmark ranges marked provisional are unverified constants; rule marks fixed scoring rules. Optional blocks appear only when asked for: shorts (with includeShorts) and rawData (with includeRawData).

The dataset has two views: overview (platform, handle, followers, tier, score, band, confidence, top red flag, data source, posts analyzed) and flat (CSV-safe columns, no nested objects, ready for Google Sheets). Flat columns include sig_er_value, sig_er_subScore, sig_vpf_*, sig_spikes_*, sig_clr_*, sig_growth_* and redFlagsText (red flags joined with " | ").

How the score works

Score = sum(weight x sub-score) / sum(weight of signals that have data), rounded. Missing signals are skipped and the remaining weights renormalized. Fewer than 3 usable signals returns INSUFFICIENT_DATA (not charged).

SignalWeightWhat it measuresHealthy range
er_vs_tier0.30Median engagement rate ((likes+comments)/followers, or /views on YouTube) against the follower tierInstagram nano 5-8%, mid 1.2-2.5% (published); other tier/platform bands are provisional
views_per_follower0.25Median views divided by followers (YouTube/TikTok)Provisional per tier
engagement_spikes0.20Posts whose likes or comments exceed 3x the median0 spikes = 100, 1 = 75, 2 = 40, 3+ = 10
comment_like_ratio0.15Total comments / total likes0.2% to 20% (provisional)
follower_growth_steps0.10Day-over-day follower jumps above 5x typical growth, from followerHistory0 steps = 100

Tiers by followers: nano < 10K, micro 10-100K, mid 100-500K, macro 500K-1M, mega >= 1M. Bands: high >= 75, medium 50-74, low 30-49, very_low < 30. Confidence: high with 4+ signals and exact counts, medium with 3 signals or approximate followers, low otherwise.

Error rows

Errors never reach the default dataset. They are written to the run's separate errors dataset (Output tab → Errors) as { "status": "error", "input": {...}, "error": { "code", "message" } }. Codes: INVALID_INPUT, INVALID_ITEM (for example a BYO profile without followers), UNSUPPORTED_DATASET_SHAPE, NOT_FOUND, INSUFFICIENT_DATA (fewer than 4 posts or fewer than 3 usable signals), RATE_LIMITED, UPSTREAM_ERROR.

Pricing

Pay per event, same events as v1:

EventPriceFires when
apify-default-dataset-item ("Audit completed")$0.015One scored profile is written to the default dataset
apify-actor-start$0.00005 (platform default)Run start

Worked examples:

RunCharge
1 YouTube channel$0.00005 + 1 x $0.015 = $0.01505
100 profiles from a scraper dataset$0.00005 + 100 x $0.015 = $1.50005
500 profiles, 20 of them fail (INSUFFICIENT_DATA)$0.00005 + 480 x $0.015 = $7.20005 (error rows are free)

You are never charged for error rows, which go to the errors dataset. The run stops cleanly when your maximum total charge is reached. This actor calls no other Store actor, so there is no hidden sub-actor cost; platform proxy and compute usage is covered by the event price.

Inputs

All fields are optional except that you must give profiles or datasetId (or a legacy field). An agent can call it with one field.

FieldTypeDefaultDescription
profilesarraynone (prefill [{"platform": "youtube", "handle": "@mkbhd"}]){platform, handle} to fetch (YouTube keyless), or add followers and posts[] (at least 4, optional followerHistory) to score your own data
datasetIdstringemptyDataset from apify/instagram-profile-scraper, clockworks/tiktok-scraper / tiktok-profile-scraper or streamers/youtube-channel-scraper
maxPostsinteger 4-5012Recent posts used. Keyless YouTube lists about 30 recent uploads (15 when it falls back to RSS)
includeShortsbooleanfalseYouTube: score Shorts separately in a shorts block
youtubeApiKeysecret stringnoneOptional BYOK: adds comment counts. An invalid key falls back to keyless with warning BYOK_INVALID
proxyConfigurationobject{"useApifyProxy": true} (datacenter)Residential proxy is used only by the unofficial Instagram/TikTok modes
tiktokUsernamesstring[] (legacy)[]v1 input, still supported: TikTok usernames for the unofficial direct mode
usernamestring (legacy)emptyv1 input, still supported: one TikTok username or profile URL
includeRawDataboolean (legacy)falsev1 option, still supported: adds rawData with the posts and follower history used

Example: one YouTube channel (no keys)

{
"profiles": [{ "platform": "youtube", "handle": "@mkbhd" }]
}

Example: your own metrics (needs at least 4 posts)

{
"profiles": [
{
"platform": "instagram",
"handle": "my_brand_account",
"followers": 8200,
"followerHistory": [
{ "date": "2026-09-01", "count": 7900 },
{ "date": "2026-09-02", "count": 7930 },
{ "date": "2026-09-03", "count": 7962 },
{ "date": "2026-09-04", "count": 7991 },
{ "date": "2026-09-05", "count": 8021 }
],
"posts": [
{ "likes": 510, "comments": 22, "timestamp": "2026-10-01" },
{ "likes": 470, "comments": 19, "timestamp": "2026-10-04" },
{ "likes": 530, "comments": 31, "timestamp": "2026-10-08" },
{ "likes": 495, "comments": 18, "timestamp": "2026-10-12" }
]
}
]
}

Recipe: score a scraper dataset without re-scraping

  1. Run apify/instagram-profile-scraper (or clockworks/tiktok-scraper, streamers/youtube-channel-scraper) on your own account.
  2. Copy the run's default dataset ID.
  3. Run this actor with:
{
"datasetId": "YOUR_DATASET_ID",
"maxPosts": 12
}

Mixed platforms in one dataset are fine. Datasets with no recognizable shape stop the run at start with UNSUPPORTED_DATASET_SHAPE and no charge. The field mapping for these three scrapers was built from their public README samples and is marked unverified: items scored from a dataset carry the warning ADAPTER_MAPPING_UNVERIFIED:<scraper>. For anything else, pass profiles[].

Legacy v1 inputs still work

tiktokUsernames, username, profiles and includeRawData are mapped onto the v2 model. A legacy-only run such as this one still returns a valid result:

{
"tiktokUsernames": ["charlidamelio"],
"includeRawData": true
}

Legacy handles are fetched with the unofficial TikTok mode (reduced signal set, confidence capped, dataSource: "tiktok_web_unofficial"). For a full five-signal score on TikTok or Instagram, pass your scraper's datasetId or your own profiles[] metrics instead.

Use with AI agents (MCP)

Agents find this actor by searching the Apify Store for "fake follower checker", "YouTube", "Instagram" or "TikTok" audience audit. Connect any MCP client (Claude, Cursor, VS Code, n8n AI Agent) to:

https://mcp.apify.com?tools=apricot_blackberry/audience-authenticity-audit

Example call-actor arguments. Only profiles (or datasetId) is needed; put the spending cap in callOptions.maxTotalChargeUsd, never in the input:

{
"actor": "apricot_blackberry/audience-authenticity-audit",
"input": { "profiles": [{ "platform": "youtube", "handle": "@mkbhd" }] },
"callOptions": { "maxTotalChargeUsd": 0.1 }
}

The run returns one item per scored profile with authenticityScore, band, confidence and signals; failures are in the run's errors dataset. Always set maxTotalChargeUsd: the run stops cleanly at that budget.

API

Run the actor and fetch the scored items in one call (maxTotalChargeUsd is a query parameter, not input):

curl -X POST "https://api.apify.com/v2/acts/apricot_blackberry~audience-authenticity-audit/run-sync-get-dataset-items?maxTotalChargeUsd=0.1" \
-H "Authorization: Bearer $APIFY_TOKEN" \
-H "content-type: application/json" \
-d '{"profiles":[{"platform":"youtube","handle":"@mkbhd"}]}'

Spreadsheet-ready export: https://api.apify.com/v2/datasets/<DATASET_ID>/items?view=flat&format=csv.

JavaScript (npm install apify-client):

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('apricot_blackberry/audience-authenticity-audit').call(
{ profiles: [{ platform: 'youtube', handle: '@mkbhd' }] },
{ maxTotalChargeUsd: 0.1 },
);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const i of items) console.log(i.handle, i.authenticityScore, i.band);

Python (pip install apify-client):

import os
from decimal import Decimal
from apify_client import ApifyClient
client = ApifyClient(token=os.environ["APIFY_TOKEN"])
run = client.actor("apricot_blackberry/audience-authenticity-audit").call(
run_input={"profiles": [{"platform": "youtube", "handle": "@mkbhd"}]},
max_total_charge_usd=Decimal("0.1"),
)
for item in client.dataset(run["defaultDatasetId"]).list_items().items:
print(item["handle"], item["authenticityScore"], item["band"])

Standby HTTP endpoint

Base URL https://apricot-blackberry--audience-authenticity-audit.apify.actor (copy the exact URL from the actor's API tab). Authenticate with Authorization: Bearer <APIFY_TOKEN>. Use it for low-latency single lookups. Each successful response is one billable "Audit completed" item.

Method and pathParametersResponse
GET /scoreplatform, handle, optional maxPosts (4-50, default 12), includeShorts=true{ "ok": true, "data": <item> }
POST /scoreJSON body: one profile {platform, handle, followers, posts[], followerHistory?}, up to 256 KB{ "ok": true, "data": <item> }
GET /healthnone{ "ok": true, "data": { "status": "healthy", ... } }

Failure returns { "ok": false, "error": { "code", "message" } } with HTTP 400, 404, 413, 422, 429 or 502.

# public YouTube channel, no keys
curl -H "Authorization: Bearer $APIFY_TOKEN" \
"https://apricot-blackberry--audience-authenticity-audit.apify.actor/score?platform=youtube&handle=@mkbhd&maxPosts=12"
# your own metrics (body up to 256 KB, one profile)
curl -X POST -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
-d '{"platform":"instagram","handle":"x","followers":8200,"posts":[{"likes":510,"comments":22,"timestamp":"2026-10-01"},{"likes":470,"comments":19,"timestamp":"2026-10-04"},{"likes":530,"comments":31,"timestamp":"2026-10-08"},{"likes":495,"comments":18,"timestamp":"2026-10-12"}]}' \
"https://apricot-blackberry--audience-authenticity-audit.apify.actor/score"

Integrations (n8n, Make, Zapier)

Use the HTTP module (Make: "Make an API key auth request", n8n: HTTP Request) with POST https://api.apify.com/v2/acts/apricot_blackberry~audience-authenticity-audit/run-sync-get-dataset-items?maxTotalChargeUsd=1, header Authorization: Bearer <your Apify token>, and the JSON body from any input example above. The response is the list of scored items; map authenticityScore, band and topRedFlag into your sheet or CRM. Zapier and Make can also use the Apify app's Run Actor action.

FAQ

Is Instagram/TikTok supported? Yes through your own scraper dataset or metrics. Direct fetching of public Instagram and TikTok profiles is an unofficial, best-effort mode (below).

Why is a signal null? The data was not public. For example keyless YouTube exposes no comment counts. The remaining weights are renormalized and confidence is lowered.

Does a low score mean fake followers or fraud? No. It means the signals warrant a closer look. The output describes risk signals and a band, never accusations about a person.

A channel hides its subscriber count? Views-based signals are used with an imputed tier, the item gets confidence: "low" and warning FOLLOWERS_HIDDEN_TIER_IMPUTED; with too few signals you get INSUFFICIENT_DATA (not charged).

Limits

  • Engagement and views-per-follower bands other than Instagram nano and mid are provisional, unverified constants; outputs mark them benchmarkSource: "provisional".
  • Keyless YouTube uses the public channel Videos tab and watch pages (dataSource: "youtube_page"), or the RSS feed when the Videos tab is unavailable ("youtube_rss", latest 15 uploads; the feed returns 404 from Apify's network); subscriber counts are rounded to 3 significant figures (followersApproximate: true). Neither source has comment counts. Keyless fetching can be rate-limited by YouTube.
  • All-Shorts channels are scored on their Shorts with warning ALL_SHORTS_CHANNEL; the benchmarks are not Shorts-specific.
  • Unofficial modes (Instagram web_profile_info, TikTok page data) are best effort, may break or be blocked at any time, use a residential proxy, are limited to 40 profiles per run with 2-5 s delays, are labelled *_web_unofficial and cap confidence at medium. TikTok direct mode only exposes followers, total likes and video count, so only the engagement signal is available and confidence is low. We scrape only public profile data and never log in.
  • Dataset adapter field mappings are unverified against real scraper output.
  • The score is a heuristic signal set, not proof of anything about a person.

Support

Built by Creator Fusion LLC (Creator Fusion Labs). Questions or issues: contact@creatorfusion.net.

If this Actor saved you time, please leave a review on its Apify Store page. Honest reviews, good or bad, are how other people decide whether to trust it, and we read every one. Found a problem? Open an issue on the Issues tab and we will fix it.