How it works
Public posts in.A report that shows its work out.
Give us a handle. We look at how the account behaves — not what it thinks — and write a likelihood score you can read. The long version, with tiers, limits, and the dated eval, lives on the methodology page.
Four steps
The same path for every account.
01
We read what’s public
The profile and recent posts on X. Nothing private. Nothing behind a login.
02
We look at behaviour
When it posts, what it repeats, how it talks to people, what the profile itself says. Not what it believes.
03
Rules add up a score
Inspectable checks, not a mystery model. Overlapping checks that tell the same story don’t stack. The number is how likely — not proof.
04
You get a report
A score, a tier, three reasons with the evidence, and everything else we checked further down. Same report on the web and in the extension.
What every scan collects
The same sample on every plan.
Posts, replies, network, AI checks
Every scan reads the account’s recent posts, opens replies under the top posts, and samples followers, following, and amplifiers. AI checks run on that sample.
What does not change
Checks that are not flagged do not lower the score.
The report names the checks
Each reason is a claim, a baseline (“most people…”), and the posts or chart that back it. If we can’t show the work, we don’t dress the number up.
Read the sample ›
What this is not
Short, so you don’t have to hunt for it.
Not a verdict on a person
A high score means the public behaviour lined up with automation or a run account. You still decide what that means.
Not politics
We do not score ideology or “truth.” Neutrality is a published page, not a slogan.
Not a secret recipe
Tiers, limits, and the dated eval are on the methodology page. The per-check catalogue is not — that’s the copyable half.
Want the long version?
Tiers, what we can and can’t see, how we check the scoring, how to dispute a report.