HopliteYC S26 landing page report.
← Back to YC S26

Score over time

90807060504030Devin (Cognition) 77 on Jul 27 (Live)Factory 70 on Apr 1 (Wayback)Factory 70 on Apr 9 (Wayback)Factory 71 on Apr 28 (Wayback)Factory 71 on May 12 (Wayback)Factory 68 on Jul 3 (Wayback)Factory 68 on Jul 23 (Wayback)Factory 68 on Jul 28 (Live)Cursor Background Agents / Cursor 75 on Jul 27 (Live)Codegen 70 on Mar 4 (Wayback)Codegen 70 on Mar 15 (Wayback)Codegen 70 on Apr 6 (Wayback)Codegen 69 on Apr 20 (Wayback)Codegen 42 on May 13 (Wayback)Codegen 53 on Jun 18 (Wayback)Codegen 72 on Aug 2 (Live)OpenAI Codex 62 on Apr 5 (Wayback)OpenAI Codex 62 on Apr 18 (Wayback)OpenAI Codex 62 on May 16 (Wayback)OpenAI Codex 54 on May 31 (Wayback)OpenAI Codex 59 on Jun 28 (Wayback)OpenAI Codex 60 on Sep 7 (Wayback)OpenAI Codex 60 on Sep 20 (Live)Hoplite 55 on Aug 4 (Wayback)Hoplite 67 on Sep 5 (Wayback)Hoplite 70 on Sep 20 (Live)Aug 4Sep 5Sep 20
Past reportsShow past reportsHide past reports
  1. Sep 20, 202670LiveOverall +3 · CTA +14 · Value props +4Viewing
  2. Sep 5, 202667Overall +12 · Value props +22 · Information hierarchy +14View →
  3. Aug 4, 202655First tracked report — strongest CTA 68, weakest Value props 42View →

Trust

58Fair

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

Hoplite
Devin (Cognition)
Trust read
Hoplite"Backed by Y Combinator" badge above the hero headline
Devin (Cognition)Embedded Devin session demo showing the task "Migrate all gradient text to #317CFF in both repos, then test" with real P
FactoryNo visible trust signal identified.
Cursor Background Agents / CursorDetailed 'Cursor Desktop' product UI showing the 'Plan Mission Control' agent workflow and Cursor CLI output
CodegenTerminal demo running 'codegen search "summarize a PR"' with real directory results
OpenAI Codeximplicit context OpenAI Codex may benefit from pre-existing familiarity among its target audience.

Value props

68Fair

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

Hoplite’s headline sellsBuild with swarms of cloud agentsHoplite turns ideas into built, tested, verified software at the speed of thought.

Working

  • Names the mechanism: swarms of cloud agents
  • 'Tested, verified' defuses AI-code distrust
  • Outcome-first: ideas become working software

Could improve

  • No numbers anywhere in the fold
  • Desktop hero shows an empty 'No threads yet' UI
  • 'Speed of thought' is an unprovable abstraction

Information hierarchy

72Fair

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

Information hierarchy annotations on the desktop capture1234
  1. Faded Start free — White text on a washed-out gradient makes the conversion CTA the least legible element in the hero.
  2. Download for Mac — Solid black styling outweighs the adjacent Start free, splitting the primary action into two equal choices.
  3. Nav Start free duplicate — A third same-label CTA in the nav competes with the hero pair instead of deferring to it.
  4. Try Pro banner — The red banner is the highest-contrast element on the page and pulls the first glance away from the headline.

Mobile web

72Fair

The phone experience as a first-class citizen.

Mobile web annotations on the mobile capture123
  1. Start free gradient button — White label on a pale grey-to-orange gradient reads low-contrast and sits half-width; duplicates the header's Start free instead of one full-bleed action.
  2. Side-by-side CTA pair — Download for Mac and Start free split into two narrow half-width buttons instead of one thumb-friendly full-width primary.
  3. Unreadable product screenshot — Desktop app screenshot shrunk to phone width renders its UI text illegibly small, adding noise without communicating anything.

CTA

78Fair

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

Hoplite
Devin (Cognition)
OpenAI Codexin-product action

Problems

  • hero 'Start free' looks disabled
  • 'Download for Mac' outweighs it
  • two hero buttons compete