Inspect HTML structure: elements that never close, close tags matching nothing, repeated ids, images without alt text, and every attribute in use.
Waiting for markup
Structural inspection, not validation against the HTML specification.
The report appears here
Structure, element and attribute inventories, and every comment with its line.
HTML Inspector runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
This is a structural inspector, and the distinction from a validator is worth being precise about rather than glossing over.
A validator checks a document against the specification's content model: which elements may appear inside which, which attributes are defined for which elements, which are required. That is a large amount of machinery and this tool does not have it. What it does instead is report what is demonstrably true about the markup you pasted — which elements never closed, which close tags match nothing open, which id appears more than once, how deep the nesting goes, which attributes are in use and on what. A document that comes back clean here may still be invalid HTML, and the page says so rather than implying a clean bill of health.
Those turn out to be the problems people actually have. An unclosed <div> shifts the rest of a page into it. A repeated id means getElementById returns the first one while CSS matches both, which produces behaviour that looks random until you find it. An <img> with no alt attribute and one with alt="" mean different things to a screen reader, and only one of them is a decision.
Optional closing tags are handled rather than reported. <li>Apple<li>Pear is correct HTML — the specification lets you omit the closing tag — so flagging it as two unclosed elements would be noise. The same applies to <p> before a block element, to table rows and cells, and to <option>.
The attribute view groups what it finds: ordinary attributes, data-*, aria-* and role, and inline event handlers. That last group is usually the interesting one, because an inline onclick is both a maintenance problem and the thing a Content Security Policy will block.
No, and it does not claim to be. A validator checks a document against the specification's content model. This reports facts about the markup — what never closed, what repeats, what is missing an attribute. Clean output here does not mean the document is valid HTML.
Because HTML lets you omit them. The closing tag for li, p, tr, td, th, option and several others is optional, and reporting correct markup as an error would make the tool useless. Anything outside that small set is reported honestly.
An id has to be unique in a document. getElementById returns only the first match while a CSS id selector matches every element with it, and label-for, aria-labelledby and fragment links all resolve to the first. The result is behaviour that looks intermittent until you find the second one.
Because alt="" is a decision: it tells a screen reader the image is decorative and should be skipped. A missing alt attribute is an absence of a decision, and the screen reader often reads the filename instead. Those are different states and the tool treats them differently.
Only the one thing it can see from the markup alone: images without an alt attribute. Real accessibility needs the rendered page — focus order, contrast, keyboard reachability. The Contrast Checker covers the arithmetic half of that.
No. HTML is not XML and this inspector follows HTML's rules, including optional closing tags and void elements, which XML does not have. For XHTML served as XML you want a genuine XML parser.
No. Everything runs in your browser and no part of the document reaches a server or analytics.