More than img src
Candidate source types distinguish ordinary images, lazy attributes, responsive sets, picture sources, social metadata and preloads.
Scan a Webpage for Image Problems
Scan a public webpage for discoverable image candidates, dimensions, formats and optimization opportunities.
Single-page scanPublic targets onlyLocal optimization
Audit discoverable image problems without pretending to render the web.
Use the scan as a focused inventory of image references visible in one page's HTML. Candidate labels point to items worth checking; they do not replace field performance data or visual review.
- Scan scope
- One public HTML document
- Candidate cap
- Up to 80 discovered URLs
- Scripts
- Never executed
Evidence over a score
There is no invented grade. The scanner exposes observations, heuristic flags and measured facts so the user keeps the final judgment.
How it works
Scanning sends the supplied public URL to the server fetch boundary. Cookies and Authorization are not forwarded, scripts are not executed and page HTML is not exposed as a generic proxy response.
Validate the target
Submit one public webpage URL. Private, loopback, link-local, reserved and metadata destinations are rejected across redirects.
Build the candidate manifest
The bounded HTML parser collects image references, then verifies supported static files three at a time and measures their bytes and natural dimensions.
Review measured results
Review the URL, supported-image count, combined source weight, estimated WebP savings, thumbnails and issue labels. All supported images start selected.
Useful for
- Inventory image references discoverable in a page's source HTML.
- Find missing dimensions and likely oversized or heavy candidates.
- Separate direct-image optimization from webpage-level discovery.
Questions about scan a webpage for image problems
What image problems does the scanner flag?
It flags missing declared dimensions, very large declared dimensions, JPEG/PNG modern-format opportunities, measured files at or above 250 KB, oversized natural dimensions and possible duplicate filename variants. These are review prompts, not a synthetic performance score.
Is this the same as a browser-rendered audit?
No. It reads a fetched HTML document without running scripts or rendering CSS. Client-inserted images, canvas content, CSS backgrounds and assets behind interaction may be absent.
Can I optimize an image after scanning?
Yes. Supported static images are measured, previewed and selected automatically after the scan. Deselect anything you do not need, then send the rest to the same worker-backed optimizer.