DEV Community

Linanxi12
Linanxi12

Posted on • Originally published at aitools-hub.xyz

I Built an AI Tools Comparison Site with a Transparent Scoring Framework

Every AI tool comparison I found was either a sponsored listicle or surface-level fluff. So I built something different.

aitools-hub.xyz


What Makes It Different

1. Public Scoring Framework

Each comparison uses 3 weighted dimensions totaling 100%. Weights are documented and adjustable per article with rationale.

Coding: Code Generation 35% / Context Understanding 35% / Debugging 30%

Image: Photorealism 40% / Prompt Adherence 35% / Artistic Style 25%

2. Consistent Structure

Every article: TL;DR -> Weighted Scores -> 3 Scenario Tests -> Detailed Comparison -> Pros & Cons -> Final Recommendation

3. All Data Sources Cited

Public benchmarks, community consensus, official docs. No fake numbers.


Current Comparisons (5 Live)

Coding Assistants

  • Cursor vs Copilot — 9.1 vs 7.6: Cursor wins on context, Copilot on ecosystem
  • Claude Opus 4.8 vs GPT-4o — 9.2 vs 8.3: Claude writes better code, GPT-4o is cheaper on API
  • Copilot vs Codeium — 8.0 vs 7.3: Copilot is better, Codeium is free and close

Image Generators

  • Midjourney v7 vs DALL-E 3 — 8.8 vs 8.3: Midjourney for beauty, DALL-E for accuracy
  • SD3 vs Midjourney — 8.2 vs 8.7: SD3 for control, Midjourney for ease

Tech Stack

  • Hugo static site generator + PaperMod theme
  • GitHub Pages for hosting
  • GitHub Actions for auto-deploy on push
  • Custom CSS for comparison tables and mobile-responsive layouts

Total monthly cost: $0 (domain is about $1/year for .xyz)


What I Learned

  1. Weighted scoring forces honesty. When your formula is public, you cannot fudge numbers.
  2. Readers value structure. Consistent format means less orientation time.
  3. Community data beats solo testing. Citing public benchmarks is more credible than claiming you ran 100 hours of tests.
  4. Hugo is underrated for content sites. Zero JS, instant loads, RSS built-in, deploys in 30 seconds.

Feedback Welcome

  • What other tool pairs should I compare?
  • Is the weighted scoring helpful or over-engineered?
  • Would you trust comparisons based on public benchmarks?

Site: https://aitools-hub.xyz
GitHub: https://github.com/Linanxi12/aitoolscompare

Top comments (0)