Your landing page’s weakest score is Trust.

Fix Trust
64Growth scoreupdated

category scores

category scoresShowHide

Trust

48Poor

Proof that the product is real, competent, and safe to try.

Salus
Zenity
Lasso Security
Prompt Security
Lakera
NeuralTrust
Trust read
Salus"PRODUCT DEMO · 59 SECONDS" section: "Watch Salus stop the wrong action before it becomes real" with concrete scenario c
ZenityGartner® named Zenity the Company to Beat in AI Agent Governance
Lasso SecurityCustomer logo row featuring Fiverr, Kaltura, Artlist, Delek US, Compugen, Kaufman Rossin, and Telit Cinterion
Prompt SecuritySentinelOne | Prompt Security co-branded header lockup
Lakera"Learning from 1M+ hackers around the world" stat in the ticker below the hero
NeuralTrustNeuralTrust has been recognized by Gartner → Read more

Mobile web

58Fair

The phone experience as a first-class citizen.

Mobile web annotations on the mobile capture1234
  1. Uncollapsed nav link row — Seven desktop nav links wrap into two rows, eating the top of the mobile fold instead of collapsing to a hamburger.
  2. Get started text link — Primary action renders as a small underlined text link at the fold's bottom edge, not a thumb-friendly full-width button.
  3. Header GET STARTED button — Desktop-proportioned pill duplicating the hero link; too small for a thumb target and stranded between logo and nav.
  4. Watch demo text link — Secondary action is plain body text with no affordance, indistinguishable from copy and barely inside the viewport.

Information hierarchy

68Fair

Does the design help or inhibit what the page is trying to say?

Salus desktop capture for Information hierarchy

What works

What to improve

CTA

68Fair

Copy, size, verb, cost-feel — does the primary CTA earn the click?

Salus

How Salus compares

Recommended rewrites

Value props

76Fair

Title, subtitle, hero — do they sell the product's true value?

Salus’s headlineAgents that act, safely.Put agents in production without losing control. Salus checks policy, evidence, and intent at runtime, returns exact fixes to the agent, and stops wrong actions before they become real.

Working

Could improve