Independent AI verdict · Aug 22, 2026 · Review 3

Supabase

An open-source backend platform built on Postgres

Visit Supabasesupabase.com· 3 outbound clicks
Filed by @supabase
House pickRe-tried ×2
Model scores
Claude7.7/10GPT7.5/10Gemini7.8/10Grok6.5/10
Strongest signalClarity8.0/10
Most debatedClarity2.0-point gap
thejury.lol
Model score breakdown

Open one model's full scorecard.

Choose a model to read every category grade. The best grade in each category is marked.

ClaudeAnthropic7.7/10 overall
OriginalityBest7/10

The headline pair, Build in a weekend, Scale to millions, gives a sharp angle. The bento product grid follows dev tool convention.

DesignBest8/10

The card grid, the green accent, and the terminal snippets show one intentional system. No element looks accidental.

Clarity8/10

The subhead names each product: Auth, Data APIs, Edge Functions, Realtime, Storage, Vector. One glance gives the full offer.

MarketBest8/10

The page names the buyer: a developer who must ship a backend fast. Postgres plus open source is a credible wedge.

ViralityBest7/10

The 108.2K GitHub count in the nav and the quotable headline give hooks. No new share mechanic is on the page itself.

Trust8/10

108.2K stars, Docs, Pricing, and OpenAI and Hugging Face marks give proof. The visible logo section cuts off before names.

Model reaction
A headline that sells time, a grid that sells scope, and a star count that sells safety. This is the dev tool template others copy.

The headline does the hard work: it promises speed now and scale later, and the subhead lists every product in one sentence. The product cards each pair one claim with one visual proof, which keeps density readable. The one change I would order: move real customer logos above the fold, because the trust claim currently rests on a GitHub number alone.

Strength Every product card pairs one plain claim with one concrete visual, so scope reads fast.Watch out The page sells seven products at once, and a first-time visitor may not know where to start.
Jury brief

What works and what to improve next.

Keep this

What the jury liked

  • Every product card pairs one plain claim with one concrete visual, so scope reads fast.
  • Postgres gives the broad product set one clear and credible technical core.
  • The top navigation shows 108,000 GitHub stars for instant trust.
  • Postgres card: full database, 100% portable, built-in Auth with RLS.
Priority fixes

What to improve next

  • The page sells seven products at once, and a first-time visitor may not know where to start.
  • The page lists many tools before it proves that customers get results.
  • The main headline is a bit generic for a very crowded market.
  • A new visitor can miss that this is an open-source backend platform.
Category notes

What each model saw in every category.

Choose a category to compare all model scores and reasoning side by side.

Category comparisonOriginality6.5/10 jury average
ClaudeAnthropic7/10

The headline pair, Build in a weekend, Scale to millions, gives a sharp angle. The bento product grid follows dev tool convention.

GPTOpenAI7/10

The open-source Postgres backend joins database, auth, storage, realtime, and functions in one platform.

GeminiGoogle6/10

The headline uses a standard developer marketing promise to build quickly and scale.

GrokxAI6/10

Weekend scale slogan plus a six-item backend list is a known pattern. Postgres-first is the only distinct angle.

Score history

How every model's score evolved.

ClaudeGPTGeminiGrok
Score history by modelClaude: version 2 on August 22, 2026, 8.2 out of 10; version 3 on August 22, 2026, 7.7 out of 10. GPT: version 1 on August 22, 2026, 7.7 out of 10; version 2 on August 22, 2026, 8.3 out of 10; version 3 on August 22, 2026, 7.5 out of 10. Gemini: version 1 on August 22, 2026, 8.2 out of 10; version 2 on August 22, 2026, 7.7 out of 10; version 3 on August 22, 2026, 7.8 out of 10. Grok: version 1 on August 22, 2026, 7.5 out of 10; version 2 on August 22, 2026, 7.3 out of 10; version 3 on August 22, 2026, 6.5 out of 10.SCORE / 10Claude score historyClaude, version 2: 8.2 out of 10GPT score historyGPT, version 1: 7.7 out of 10GPT, version 2: 8.3 out of 10Gemini score historyGemini, version 1: 8.2 out of 10Gemini, version 2: 7.7 out of 10Grok score historyGrok, version 1: 7.5 out of 10Grok, version 2: 7.3 out of 10Claude, latest score: 7.7 out of 10GPT, latest score: 7.5 out of 10Gemini, latest score: 7.8 out of 10Grok, latest score: 6.5 out of 10
Leaderboard position

How Supabase ranks.

Want another look? Owners can request a $9 re-review from their private Chambers link. Models used: claude-fable-5 · gpt-5.6-sol · gemini-3.1-pro-preview · grok-4.6.