Instagram Fake Followers Checker

Find out how much of an Instagram audience is real. Paste a handle like nasa, @natgeo, or a profile URL, and get a fake-follower percentage, an audience quality grade, an estimated real reach — and a scored row for every follower that was checked, with the exact reasons it was flagged.

---

Why Use This Checker?

• A fake-follower percentage backed by evidence, not a black-box number
• One scored row per sampled follower — post count, follow ratio, bio, avatar, handle pattern, and the flags that fired
• Audience quality score out of 100 and an A–F grade you can put in a report
• Estimated real follower count, extrapolated from the sample
• Deep mode looks up every sampled follower's own profile, which is where the real bot signal lives
• Fast mode screens from the follower list alone, roughly 50x cheaper, for scanning many accounts at once
• Audit any public account — an influencer you are about to pay, a competitor, or your own
• No login, no cookies, no session tokens — nothing to get your account flagged

---

Overview

The Instagram Fake Followers Checker is built for brands vetting influencers before a paid collaboration, agencies reporting on campaign quality, and creators auditing their own audience after a follower spike.

It produces two row shapes in one dataset, labelled by rowType:

• summary — one row per audited account: the counts, the percentage, the grade, and the flags that fired most often.
• follower — one row per sampled follower: their public numbers, their bot score, their verdict, and the flag list behind it.

Nothing is hidden behind a proprietary index. Every verdict can be re-derived from the fields in its own row.

---

Supported Inputs

Input types

Input	Example	What happens
Handle	nasa	Audited directly
@-prefixed handle	@natgeo	The @ is stripped
Profile URL	https://www.instagram.com/apify/	Handle extracted from the URL

Copy-pasteable input

{
  "usernames": ["nasa", "@natgeo", "https://www.instagram.com/apify/"],
  "followersSampleSize": 200,
  "analysisDepth": "deep",
  "includeFollowerRows": true
}

Unsupported inputs

• Private accounts — Instagram does not serve their follower list to non-followers, so there is nothing to sample. These return an error row.
• Numeric user IDs — pass the handle instead.
• Post or reel URLs — this actor audits accounts, not posts.

---

Use Cases

Influencer vetting before a paid deal. Run the creator's handle at a 500-follower sample and read the grade. A B or better is a normal organic audience; a D or F on an account with a big follower number is the classic bought-audience shape.

Campaign reporting. Audit every creator in a campaign in one run and put the fake percentages side by side in the client deck.

Self-audit after a follower spike. A sudden jump in followers with no matching jump in engagement usually shows up here as a wave of zero-post accounts.

Competitor benchmarking. Compare the audience quality of accounts in your niche, not just their follower counts.

Lead qualification. Filter out prospects whose apparent reach is manufactured before spending sales time on them.

---

How It Works

1. Each handle is resolved to a profile — follower count, following count, post count, verified and private status.
2. Followers are sampled page by page (about 50 per request) up to followersSampleSize.
3. In deep mode, every sampled follower's own profile is looked up: post count, follower and following counts, bio, external link.
4. Each follower is scored on additive signals — no posts, almost no followers, lopsided follow ratio, mass following, empty bio, default avatar, no display name, auto-generated handle patterns. Verified accounts short-circuit to a clean score.
5. A score of 60+ is likely_fake, 35–59 is suspicious, below 35 is likely_real. The summary counts each bucket, averages the scores into a quality score out of 100, and grades it A–F.

Weights are deliberately tuned so no single weak signal can condemn an account. A real person with an empty bio and no display name lands at suspicious at worst; it takes the full bot shape — no posts, no followers, hundreds followed — to reach likely_fake.

---

Input Configuration

Field	Type	Default	Description
usernames	array	—	Public accounts to audit. Required.
followersSampleSize	integer	200	Followers to analyse per account (10–5000).
analysisDepth	select	deep	deep looks up each follower's profile; fast screens from the list only.
includeFollowerRows	boolean	true	Return the evidence row per follower alongside the summary.

Example — vetting one creator thoroughly

{
  "usernames": ["@somecreator"],
  "followersSampleSize": 1000,
  "analysisDepth": "deep",
  "includeFollowerRows": true
}

Example — cheap screen across a shortlist

{
  "usernames": ["creator_one", "creator_two", "creator_three"],
  "followersSampleSize": 300,
  "analysisDepth": "fast",
  "includeFollowerRows": false
}

---

Output Overview

One summary row per audited account, plus one follower row per sampled follower when includeFollowerRows is on.

Accounts that could not be audited — not found, private, or with no follower list served — are returned as error rows carrying username, error, and message. These are never charged.

---

Output Samples

Real summary row, produced from usernames: ["nasa"] with a 20-follower sample:

{
  "rowType": "summary",
  "username": "nasa",
  "userId": "528817151",
  "fullName": "NASA",
  "isVerified": true,
  "isPrivate": false,
  "followerCount": 104428024,
  "followingCount": 92,
  "mediaCount": 4890,
  "analysisDepth": "deep",
  "sampledFollowers": 20,
  "likelyFake": 4,
  "suspicious": 7,
  "likelyReal": 9,
  "fakePercent": 20,
  "audienceQualityScore": 67,
  "grade": "C",
  "estimatedRealFollowers": 83542419,
  "topFlags": [
    { "flag": "no_posts", "count": 11 },
    { "flag": "empty_bio", "count": 9 },
    { "flag": "almost_no_followers", "count": 6 }
  ],
  "profileUrl": "https://www.instagram.com/nasa/"
}

Real follower row from the same run — a flagged account. The handle, ID and display name are redacted here; your dataset carries the real ones:

{
  "rowType": "follower",
  "auditedUsername": "nasa",
  "username": "<redacted>",
  "userId": "<redacted>",
  "fullName": "<redacted>",
  "isPrivate": false,
  "isVerified": false,
  "hasDefaultAvatar": false,
  "followerCount": 4,
  "followingCount": 512,
  "mediaCount": 0,
  "followingToFollowerRatio": 128,
  "biography": "",
  "externalUrl": null,
  "botScore": 75,
  "verdict": "likely_fake",
  "flags": ["no_posts", "almost_no_followers", "lopsided_follow_ratio", "empty_bio"],
  "profileUrl": "https://www.instagram.com/<redacted>/"
}

Error row for a private account:

{
  "username": "someprivateaccount",
  "error": "private_account",
  "message": "@someprivateaccount is private — Instagram does not serve its follower list, so its audience cannot be audited."
}

---

Key Output Fields

Summary rows username, userId, fullName, isVerified, isPrivate, followerCount, followingCount, mediaCount, analysisDepth, sampledFollowers, likelyFake, suspicious, likelyReal, fakePercent, audienceQualityScore, grade, estimatedRealFollowers, topFlags, profileUrl

Follower rows auditedUsername, username, userId, fullName, isPrivate, isVerified, hasDefaultAvatar, followerCount, followingCount, mediaCount, followingToFollowerRatio, biography, externalUrl, botScore, verdict, flags, profileUrl

Flags you may see no_posts, almost_no_followers, lopsided_follow_ratio, mass_following, empty_bio, private_and_empty, default_avatar, no_display_name, digit_suffix, long_digit_suffix, mostly_digits, placeholder_handle, no_vowels, promo_handle, verified_account

---

FAQ

How accurate is the fake percentage? It is a sample estimate from public signals, not a verdict from Instagram. A 200-follower sample places the percentage within a few points for most accounts; raising the sample tightens it. Treat a grade difference (B vs. E) as meaningful and a two-point difference as noise.

Why is a big verified brand not graded A? Large accounts accumulate dormant and abandoned followers over years — accounts with no posts, no bio, and almost no followers of their own. They look identical to purchased accounts from the outside, and they dilute reach the same way. A C on a hundred-million-follower account is normal.

What is the difference between deep and fast? Deep looks up each sampled follower's own profile, so it can see post counts and follow ratios — the signals that actually separate bots from lurkers. Fast only sees what the follower list carries (handle, display name, avatar) and costs about one request per 50 followers instead of one per follower.

Fast systematically under-reports, and by a lot: auditing the same account both ways returned 20% fake / grade C on deep and 1.7% fake / grade A on fast. That is not a bug in either mode — the bought-account signature is a zero-post profile following hundreds of accounts, and fast mode cannot see either number. Use fast to rank a shortlist against each other, deep for any figure you intend to quote or act on.

Can I audit a private account? No. Instagram does not serve a private account's follower list to anyone who does not follow it, so there is nothing to sample. The run returns an error row saying so.

Why did I get fewer sampled followers than I asked for? Instagram limits how deep the follower list goes, especially on very large accounts. The sampledFollowers field always reports what was actually analysed.

Does turning off follower rows make a run cheaper? No. Followers are charged per follower analysed, not per row returned, because a deep audit spends an upstream lookup on each one whether or not you keep its row. The toggle controls how much evidence lands in your dataset, not the bill. To spend less, lower the sample size or switch to fast.

Does a suspicious verdict mean the account is fake? No. It means the account shows some of the shape without enough of it to call. Dormant real accounts land there routinely. Only likelyFake feeds the fakePercent figure.

Can I check which specific followers are fake? Yes — that is what the follower rows are. Sort by botScore descending and read the flags column.

---

Support

Found a bug or need a field that is not there? Open an issue on the actor's Issues tab and it will be looked at.

Email: muhamed.didovic@gmail.com

---

Additional Services

Need a custom scraper, a tailored export format, or an integration into your own pipeline? Custom builds are available — get in touch at muhamed.didovic@gmail.com.

---

Explore More Scrapers

• Instagram Followers & Following Scraper — full follower and following lists
• Instagram Profile Scraper — profiles, posts and comments from any public account
• Instagram Post Likers Scraper — who liked any public post or reel
• Instagram Comments Scraper — full comment threads with replies
• Instagram Similar Profiles & Lookalike Finder — accounts similar to any profile
• Instagram Hashtag Scraper — top posts and reels from any public hashtag
• Instagram Location Scraper — place cards and geotagged posts

Browse the full collection at apify.com/memo23.

---

🤖 For AI Agents & LLM Apps

Compact reference for AI agents calling this actor via the Apify MCP server or the Apify API (actor: memo23/instagram-fake-followers-checker).

Purpose: Samples a public Instagram account's followers, scores each for bot-like signals, and returns a fake-follower percentage with an audience quality grade.

Minimal input:

{ "usernames": ["nasa"], "followersSampleSize": 200, "analysisDepth": "deep" }

Output: two row shapes in one dataset, distinguished by rowType. summary rows carry sampledFollowers, likelyFake, suspicious, likelyReal, fakePercent, audienceQualityScore, grade, estimatedRealFollowers, topFlags. follower rows carry the sampled account's public numbers plus botScore, verdict and flags. Unauditable accounts return {username, error, message} instead.

Behaviors an agent should know:

• analysisDepth: "deep" costs roughly one upstream request per sampled follower; "fast" costs one per 50. Pick deep for decisions, fast for ranking.
• Private accounts always return error: "private_account" — check before scheduling a run over a list.
• sampledFollowers may be lower than followersSampleSize; Instagram caps follower-list depth on large accounts.
• fakePercent counts only likely_fake verdicts. suspicious is a separate bucket and is not included.
• estimatedRealFollowers extrapolates the sample across the full follower count — it inherits the sample's margin of error.
• Billing: one profile event per summary row, one follower event per follower ANALYSED. includeFollowerRows: false suppresses the rows but not the charge — a deep audit looks each follower up either way. Error rows are never charged.
• No login is used, so no authenticated-only fields are available.

---

⚠️ Disclaimer

This Actor is an independent tool and is not affiliated with, endorsed by, or sponsored by Meta Platforms, Inc. or any of its subsidiaries. All trademarks mentioned are the property of their respective owners.

The scraper accesses only publicly available Instagram profile and follower data — no authenticated endpoints, paid features, or content behind the instagram.com login wall. Scores and verdicts are statistical estimates derived from public signals, not statements of fact about any individual account, and should not be used as the sole basis for consequential decisions about a person. Users are responsible for ensuring their use complies with instagram.com's Terms of Service, applicable data-protection law (GDPR, CCPA, etc.), and any contractual obligations of their own organization.

---

SEO Keywords

instagram fake followers checker, instagram bot detector, fake follower audit, instagram audience quality, influencer vetting tool, check fake followers instagram, instagram engagement audit, bot follower detection, influencer fraud detection, instagram follower analysis, audience authenticity score, instagram account audit, social media fraud check, creator vetting data, instagram follower quality, instagram.com scraper, Apify instagram, no login instagram scraper