Quick findings, instantly
Click the Verity icon. The article text is extracted and a small model running on your own device flags likely loaded language and unattributed claims within a second. Nothing leaves your browser for this step.
Chrome extension · Junction 2026
See the spin before it sees you.
Verity scans the article you’re reading, highlights manipulative language and questionable claims, and explains in plain words which technique is being used, so you learn why a sentence is trying to steer you, not just that it is.
Free · Works in Chrome, Edge and Brave · Installed manually while it awaits the Chrome Web Store
Hero video
16:10 · assets/hero.mp4
How it works
One click, three layers of checking.
Smart analysis, explained
An AI analysis reviews the article and names the specific technique behind each flagged passage, with a short explanation in plain language and a confidence level. Click a finding to jump to it on the page.
Fact-check with sources
For factual claims, ask Verity to check them against the web. You get a verdict (supported, contradicted, disputed or couldn’t verify) with links to the sources it used.
What it detects
Techniques, not opinions.
Verity judges how something is written, not which side it takes. Arguing a point is fine; manipulating the reader is what gets flagged.
Loaded language
Emotionally charged wording used where neutral wording was available: “shameless cash grab”, “invasion”.
Propaganda techniques
Name calling, appeals to fear, flag-waving, slogans, bandwagon, casting doubt without evidence.
Logical fallacies
Straw man, whataboutism, false dilemma, causal oversimplification, slippery slope.
Unsupported claims
Numbers, events or causes stated as fact with no source: “experts say”, “studies show”.
Missing context
Technically true but misleading: cherry-picked statistics, missing baselines, one-sided framing.
Misinformation
Only when web sources contradict a claim. Without evidence either way, it stays an unsupported claim.
Honest by design
Built to avoid crying wolf.
Fewer, better flags
The analysis is told to flag only clear instances and return nothing for neutral, well-sourced writing. Every flag must quote the exact words that triggered it; flags that quote text not in the article are discarded automatically.
Reported speech isn’t the article’s spin
Quoting someone who uses loaded language is reporting. Verity only flags rhetoric inside quotes when the article itself adopts or amplifies it.
“Misinformation” needs evidence
A claim is only called misinformation when sources found on the web contradict it, and those sources are shown to you. Everything else is labelled as unsupported or disputed.
Quick vs smart, clearly labelled
On-device Quick findings are fast but less accurate, and are marked as such. Confidence is shown for every finding, and AI analysis can be wrong: treat Verity as a second pair of eyes, not a verdict.
Measured accuracy
Precision, recall and false-positive rate on a labelled test set and on neutral control articles.
Install
Up and running in a minute.
Verity isn’t on the Chrome Web Store yet, so it’s installed as an unpacked extension. It works in Chrome, Edge, Brave and other Chromium browsers.
Download verity-extension.zip-
1
Unzip the download
Extract
verity-extension.zipto a folder you’ll keep, for example in Documents. -
2
Open the extensions page
Go to
chrome://extensions(oredge://extensions) and switch on Developer mode in the top right. -
3
Load the folder
Click Load unpacked and choose the folder you unzipped.
-
4
Pin it and scan
Pin Verity from the puzzle-piece menu, open any news article and click the Verity icon.