Channel Authenticity Check

Vet a YouTube channel for fake views, bot comments, and non-organic engagement before you book a sponsorship. Returns a 0–100 authenticity score with ranked red-flag findings — calibrated on real bought-view and comment-farm investigations.

Use in Claude Code Set up your agent

Know whether a channel’s audience is real — before you spend a dollar on it.

Why the score holds up

A score is only as good as what it’s measured against. This one is checked against Firebolt’s 8.5B video snapshots — the real view curves of the 1,302,318 channels we track — so “this spike isn’t organic” means it doesn’t look like real channels in the same niche, not a hunch. The comment read runs over the transcripts and engagement already sitting in Elasticsearch, and the peer baseline is built from actual niche neighbors, not a generic benchmark.

We’ve also seen the failures up close. Across eight years and 26,142 sold deals, we’ve run the post-mortems on sponsorships that underdelivered — the bought-view campaigns, the comment farms — and that history is what the red-flag rules are calibrated on. Point the skill at a deal you already ran and it audits the views you paid for against that same record. It’s the data moat aimed at one question: is this audience real?

Try this prompt
> Vet this YouTube channel for fake views and bot comments before I sponsor it: youtube.com/@mkbhd

Hand Claude a channel handle, URL, ID, or name. It runs the full audit and hands back an authenticity score.

To Claude

How it works

  1. Collect
    Pull the channel's recent videos, view curves, engagement, and comments.
  2. Compare
    Measure them against a niche-matched peer baseline to surface anomalies.
  3. Read comments
    Classify the audience: genuinely engaged, or templated, bot-like, and spam.
  4. Score
    Return a 0–100 authenticity score with ranked red flags and a clear verdict.