DiscoReady
Free tools▾
🔎 Profiler — Available 🔬 Title Lab — Available 🖼️ Image Validator — Available 🧬 Schema Auditor — Available ⚡ 1-min Quick Audit — Available
✍️ Title generator · Soon
Discover Audit SOON
Solutions▾
Available
🤝 Discover Consulting
Soon
🎓 Training & Masterclass
🔌 Plugins & Addons
🛠️ Site management
🏪 Marketplace
Resources▾
📘 Discover Essentials Ebook — Available
📚 Free guides · Soon
📊 Case studies · Soon
📖 Discover glossary · Soon
📩 Monthly newsletter · Soon
Blog
🇫🇷 🇬🇧
Get started
Home › Free tools › Image Validator
Image Validator · AI · free

Will your hero image stop the scroll on Discover?

Submit an image URL — we audit it in 5 seconds: 8 technical checks (size, ratio, format, compression, MP) + AI visual analysis (subject clarity, brightness punch, color richness, Discover editorial fit, thumb-stopping power). Verdict + actionable fixes.

8 technical checksAI visual analysisVerdict + actionable fixes
JPG, PNG, WebP
Lets the AI score title ↔ image alignment

✨ Tip: with a title provided, the AI also rates whether the image fulfills the title's promise — a major Discover signal (mismatch = perceived clickbait = algo penalty).

No image is stored. Validation runs in memory and is never persisted.

Method

How it works

1

Download

We fetch the image from the URL you provide (HTTPS only, redirects followed, capped at 8 MB for safety).

2

Technical audit

8 binary checks: width, height, ratio, format, weight, megapixels, compression, Content-Type. Each pass / warn / fail with detail.

3

AI visual analysis

Claude AI rates 6 dimensions: subject clarity, brightness punch, color richness, freedom from text overlay, emotional pull, Discover editorial fit.

4

Click-stopping verdict

Composite 0-100 score + verdict (strong / decent / weak) + 2-4 concrete fixes to reposition the image.

✨ FAQ

Frequently asked questions

How does the AI analysis work?

We send your image to Anthropic's Claude vision model with an expert prompt that asks for 0-100 ratings on 6 axes (subject clarity, brightness punch, color richness, freedom from text overlay, emotional pull, Discover fit) + a click-stopping score. The structured JSON response feeds the verdict and the recommendations.

What dimensions does Google Discover require?

Discover prioritizes images ≥ 1200 px wide with max-image-preview:large in the robots meta. Below that, the image gets a mini thumbnail or doesn't surface at all.

Which ratio is recommended?

16:9 or 4:3 landscape, sometimes 1.91:1 (OG). 1:1 (square) and portrait are less well supported in the Discover feed. We tolerate ±5% around target ratios.

Which format should I pick?

WebP (best quality/weight ratio), then JPEG. PNG works but stays heavier. Avoid GIF, BMP, AVIF (still poorly supported by Discover).

Why is weight so critical?

A hero image often weighs more than the rest of the page combined. Past 1.5 MB, your LCP and therefore your Core Web Vitals get penalized — and Discover watches those.

Does the AI store my images?

No. The image is processed in memory in a serverless function, sent to Anthropic in base64 for visual analysis, then discarded as soon as the response comes back. No retention on DiscoReady's side or Anthropic's side (default zero-retention policy).

Is the service really free?

Yes — 3 free validations per day, no sign-up. For batch audits or API access, the Pro plan starts on quote.

✨ Pricing

Simple. Fair.

Start free, scale up when you need to.

Free
$0
3 audits / day
  • ✓ 8 technical checks
  • ✓ AI visual analysis (6 axes)
  • ✓ Click-stopping verdict
  • ✓ Actionable recommendations
  • — No batch audit or API
Audit my image
Pro

Let's talk about your needs

Batch audits (CSV), API, CMS integration: we'll shape the offer with you.

Contact us

Before you publish, check your headline too

A perfect image won't save a weak title. Title Lab scores your H1 against 500,000 Discover articles.

🔬 Run the Title Lab →