Your feed, made transparent.

— already on the list · free while in beta

Verity's actual report on a post: AI likelihood 72 and authenticity 52 rings, with insights flagging manufactured engagement, AI signals in text and media, humanized text, and thin replies

Your feed is an auction you can't see. A third of what trends was bought, boosted, or written by a model. The platform knows exactly which posts paid for their reach — it just doesn't tell you.

Verity is a Chrome extension that puts the receipt on every post: what was paid for, what was manufactured, and what's actually a person talking.

Every post gets a receipt.

Four labels, applied in about a second — right where you already scroll.

Verity's AI tab: every sentence scored separately, with AI-laundered text flagged and its techniques named

Reads which sentences a model wrote. Text is scored line by line — partial edits and humanized rewrites show as hot islands.

Verity's coordination tab: each account driving the numbers carries a bot score and named flags

Farmed numbers have a shape. Every account behind the count gets a bot score and named flags — bought farms separate from organic pods.

Verity's audience tab: projected age, region, and interest of the accounts a post actually reached

Knows who it actually reached. Age, region, and interest projections surface audiences that don't match what the post claims.

Labeled, not accused.

Paid reach is a label, not a verdict. Official X ads are identified — and never scored as fake.

Verity's AI tab: the sentence-level heatmap, named AI signals, and the writing metrics behind the score

AI-written copy gets called out. Cadence, burstiness, and structure vs model-output patterns.

Verity's engagement tab: per-action inorganic estimates with platform ratio bands

Ratios over vibes. Each metric is compared to the organic band, not a global threshold.

A promoted X post inside Verity, tagged Paid ad with paid reach explicitly excluded from fraud scoring

Ads identified, never confused. Promoted posts get the Ad tag — and are excluded from fraud scoring.

02 — Under the hood

A forensic read,
not a vibe check.

Every post decomposes into signal groups — text forensics, engagement bands, amplifier history, coordination forensics — fused into one verdict you can audit line by line.

Verity's engagement tab: every action scored separately, with the estimated inorganic share and the evidence behind it
Likes per vieworganic

0.019 — inside the organic band

Reposts per likeorganic

0.137 — inside the organic band

Replies per likethin

0.009 — replies suspiciously thin for reach

Views vs followerspaid

Reach outruns the base — attributed to the ad, not fraud

0signal groups per post
0amplifiers cross-checked
0boost-history window
0from click to verdict

What X shows you.
What Verity shows you.

X shows Verity
A promoted post bought its reach tiny "Ad" Ad · paid — official
Likes bought to fake traction nothing Bought engagement · %
Text written by a model nothing AI-written · %
The same accounts boosting every post nothing Pod · n amplifiers
A genuinely viral post indistinguishable organic · verified

03 — How it works

Lives where
you already scroll.

  1. 1

    Install once

    A Chrome extension, not another app. No account needed to see labels — sign in only to save reports.

  2. 2

    Browse as usual

    A small shield appears under every post. Official ads are tagged in-line so paid reach is never mistaken for fraud.

  3. 3

    Click for the autopsy

    One tap opens the full report: per-metric verdicts, amplifier forensics, and a coordination map — in about a second.

04 — Early access

The feed is loud.
The labels shouldn't be.

We're opening Verity in small waves. Leave your name, email, and number — we'll text or email you an install link when your spot opens.

Free while in beta · — already on the list