MAGNET · the product
One product. Scan, fix, prove the score moved.
MAGNET scores how readable your site is to the AI agents that recommend before a human ever lands, ships the fixes, and proves your readiness score moved. Not an audit you file away, a product you run.
Checks46Dimensions7Scan~30sLanguagesEN / AR
The loop
Scan, fix, verify, repeat.
Agent-readiness is not a one-time score. It moves as your site and the engines change, so MAGNET runs as a loop.
01
Scan
We score your site against the rubric AI agents reward, across 7 weighted dimensions, in about 30 seconds.
02
Fix
Each finding comes with a concrete fix artifact. Apply the highest-impact ones first from a ranked backlog.
03
Verify
Re-scan to confirm the fix landed and watch the score move. The loop repeats as the agentic web shifts.
What you get
A product, not a PDF you file away.
Score with coverage
A 0–100 agent-readiness score with the measured, pending and N/A map shown in full. Coverage is never hidden behind a tier.
Prioritized fix backlog
Every finding ranked by impact, each with a one-click fix artifact you can apply, not just a vague recommendation.
Monitoring and alerts
Re-scans on a schedule with email alerts when your readiness moves, so a regression is caught before your buyers' agents see it.
Before and after proof
Verified deltas after each fix ships. The readiness-score lift is measured against a baseline, not asserted.
A real dashboard
Score, trend, fixes and monitoring in one console. Dark by default, fully bilingual, mirrored for RTL.
Bilingual PDF reports
Client-ready reports in English and Arabic, generated from the same scored data the dashboard shows.
One engine, many surfaces
Meet your buyers where they already are.
The same scoring engine runs everywhere your site is discovered, from a public scan to your store platform.
Free public scanner
Score any URL in seconds, no signup to see the map.
Store connectors
Salla, Zid and Shopify apps that score and fix in-store. Coming soon.
Chrome extension
Check any page you're on against the rubric. Coming soon.
MCP server
Let agents run a free scan through the Model Context Protocol. Coming soon.
Is it for you
Built for some sites, not all.
Run MAGNET if
- Buyers research you through ChatGPT, Gemini or Perplexity.
- You sell in EN and AR and need both read correctly.
- You want fixes shipped and verified, not just flagged.
Skip it if
- Your site is offline or behind a full login wall.
- You only want a one-time opinion, not a running loop.
Get started
See your score in 30 seconds.
Run the free scan, see every finding, then pick a plan when you want the fixes shipped and monitored.