Fake followers are cheap to buy and easy to count. They're terrible at acting like people. Here's what gives them away.
We compare likes and comments against what real creators of the same size get on the same platform. A 2M account with 400 likes means most followers never show up. Bought likes without comments, or comments without likes, get flagged too.
Organic growth is a ramp. Bought growth is a staircase: +40% in a month with no viral post. Then a cliff, when the platform sweeps out fake accounts. We look for both.
Every username gets a bot score: long number strings, random letters, spam keywords, default-style handles, and batches of accounts with the same name pattern. With a CSV we also check posts, profile pics and following ratios.
Specific questions and inside jokes are real people. Emoji-only, "nice pic 🔥" on repeat, exact duplicates and "DM us to collab" are pods and bots.
Views compared to followers. Real audiences watch. If a tiny fraction ever sees a video, or likes outnumber realistic viewer behavior, something's off.
Each signal gets a 0–100 score, weighted and combined. Very weak engagement caps the total, because no amount of nice comments makes up for an audience that isn't there.
Three verdicts. No "it's complicated."
We only score what you paste in. We can't see private analytics like audience location or story views, and we can't prove intent. Viral moments and giveaways can look odd too. Treat a low score as your cue to ask the creator for their insights screenshots before you sign. Real creators send them in about a minute. Fake ones suddenly get very busy.