Evergreen LabsEvergreen Labs

Introducing Credibility Analyzer

· Billy Lo

TL;DR

Highlight any text. Right-click to analyze credibility. You get a score and explanation on the spot.

It runs on browsers (Chrome, Safari, Firefox). It also can run natively on iOS, Android and on macOS (as Share extension).

Why?

Most people know they should check questionable claims. Few do it in the moment though.

Not because we don’t care — because the usual path takes too much effort. Open a new tab. Create a prompt to check for truthfulness. Skim the responses. Cross-check sources. By then the article has scrolled on, the group chat has moved, and the claim sits unchallenged in the back of your head.

That friction is the enemy of truth. If verification costs five minutes, we skip it. If it costs two clicks, more people will do it whenever they want.

Enter Credibility Analyzer

Credibility Analyzer exists for one job: let you fact-check selected text without leaving the page.

  1. Highlight the sentence (or paragraph) that feels off.
  2. Right-click → Analyze credibility (or use Share to Credibility Analyzer on iOS/Android/Mac).

You get a 1–10 credibility / factualness score, a short justification, and claim-by-claim bullet points for easy digestion. Citations are shown too. Then you decide what to believe — with a little more context than vibes.

For even faster analysis, just share the entire page or use the toolbar icon for 1-click full-page analysis.

Why “everywhere” matters

A fact-check habit only sticks if it lives where you actually browse.

So Credibility Analyzer runs on:

  • iOS, Android and Mac — you can fact-check anything from any apps.
  • Chrome, Firefox, and Safari — you can fact-check anything on a web page.

Same idea on each surface: the check is next to the text, not buried in another app you have to remember to open later.

How we keep the trust model simple

  • No account from us, no ads SDK, and we don’t run a backend that stores your page content.
  • On demand only — nothing runs until you ask.
  • Bring your own API keys for Gemini Flash, Claude Sonnet, or GPT-6 Sol — analysis uses the provider you choose.
  • AI is not always right, but it’s a fast and decent second opinion. With a well-crafted prompt, I found Credibility Analyzer to be quite useful.

Install

Get it from evergreen-labs.org/products/credibility-analyzer, or go straight to each store:

One-time setup: pick a model, paste an API key for that provider, then try it on a claim you already distrust. Once you build the muscle memory, it can be a seamless part of your information diet.

Closing

I found it especially useful for checking posts on Twitter and Facebook, as many posts on our feeds are from unknown sources. They only look real; there is often no real human behind them.