Detect bots, engagement pods, and purchased followers across TikTok, Instagram, YouTube, X, and more. Prove your audience is real — or fix the problem before a brand deal falls through.
Audience authenticity is the percentage of a creator's followers that are real, active humans — not bots, engagement pods, or purchased accounts. CreatorScore measures it via the Authenticity agent, which blends follower pattern analysis, like-to-comment anomalies, engagement pod clustering, and growth velocity into a single 0-100 score. It contributes 20% to the overall CreatorScore and is the first signal brands check during campaign vetting.
The Authenticity agent blends four independent fraud signals into one score.
Profile pattern matching, engagement velocity anomalies, and network clustering to flag accounts that are not real people. We cross-reference follower-growth spikes with content quality drift to detect batch-purchased followers.
Identifies coordinated like/comment groups — typically 20-200 accounts that mutually engage on each other's posts within the first hour of publication. The CreatorScore algorithm measures pod_size and pod_avg_size signals to apply graduated penalties.
Sudden follower spikes inconsistent with content output, geographic mismatches (audience in countries the creator has never targeted), and follower-to-engagement ratio anomalies. A 1M account with 0.3% engagement almost always has purchased followers.
How the Authenticity agent maps raw signals to a brand-facing tier.
The CreatorScore Authenticity agent blends four independent signals: follower pattern analysis, like-to-comment anomaly detection (like_comment_anomaly), engagement pod clustering (pod_size and pod_avg_size), and follower growth velocity. It runs on every creator in the database and contributes 20% to the overall 1-100 CreatorScore. Accuracy is validated against known botted accounts in our training set and continuously improved via new signal wiring.
Five primary signals: (1) follower velocity spikes disconnected from content performance, (2) like-to-comment ratio anomalies (real audiences comment at predictable rates), (3) engagement pod size — typically 20-200 coordinated accounts, (4) audience geographic consistency (bot farms cluster in specific countries), and (5) follower-to-engagement ratio vs platform benchmark. The agent outputs an authenticity percentage from 0-100%.
Yes. Brands running campaign vetting through CreatorScore see the Authenticity score as part of the 7-agent breakdown on the creator profile. If your audience is clean, this is a major selling point and a defensible premium justifier. If the score is flagged, fixing it should be your top priority before pitching high-value brands.
Natural follower decay, inactive accounts, and spam accounts all get filtered in the scoring. You are only penalized for signals of deliberate fraud — pods, purchased spikes, botnet clustering. The scoring is designed to be fair to legitimate creators who have collected some inactive followers over time. The Authenticity agent distinguishes between "inactive but real" and "bot or purchased".
Yes, but it takes 30-60 days. Stop using engagement pods immediately. Avoid any third-party follower growth service. Run Instagram's built-in "remove inactive followers" cleanup. Then focus on natural engagement growth — your ratio will rebuild as real audience activity catches up. The CreatorScore monthly rescore will pick up the improvement on the next cycle.
CreatorScore is built into a 7-agent brand safety engine, not just a fake follower tool. You get the Authenticity score plus Content Risk, Brand Safety, Audience Quality, Sentiment, Community Trust, and ROI Prediction in one unified 1-100 score. HypeAuditor and Modash are dedicated audience analytics tools; CreatorScore is a full influencer vetting platform that includes audience authenticity as one dimension.