Why not just ask a chatbot to review my HTML?

Because a language model reading markup cold will give an impression, and can miss or invent issues. This tool runs a real static scan against WCAG rules and reports measured findings, pointing at the actual elements, so the answer is grounded in what the code contains rather than a guess.

Which issues does it catch?

It covers common WCAG failures that static analysis can detect: missing alt text, unlabeled form fields and controls, low color contrast, skipped heading levels, and similar structural problems. Each finding is tied to a rule and the element that breaks it.

Does the scan cost credits?

The scan itself is free. Running it draws nothing from your balance; you only pay the normal model-token cost of the answer Zeplik writes around it, exactly like any other message, and the cost is shown before you send.

Can it check a live URL instead of pasted HTML?

The scan works on the HTML you paste, not a live rendered page. For a deployed page, copy its markup into the chat, or pair this with a full browser-based audit tool that evaluates the rendered result. Zeplik can also read a page's text if you give it the link.

Does a clean scan mean my page is fully accessible?

No. Static analysis catches many issues but not everything, and some things, like whether alt text is actually meaningful or the tab order makes sense, need a human or assistive-technology test. Treat a clean scan as a strong start, not a compliance certificate.

Which AI model does the scan?

The scan is a built-in capability that works the same regardless of model. The explanation around it is written by whichever of Zeplik's frontier chat models you pick, and you can switch models at any time, including mid-conversation.

Accessibility checker

Paste your HTML and Zeplik runs a real scan for WCAG accessibility issues, then answers from the measured findings. Not a general impression of your markup: it points at the exact elements that break a rule, from missing alt text and unlabeled form fields to color contrast that falls short of AA, and explains the fix for each. A clean scan is a strong start, and Zeplik is honest that it is not a full compliance certificate.

Open the design assistant

The scan is free; you pay only the normal cost of the answer, shown before you send.

How it works

Paste your HTML
Drop the markup for a component or page into the chat and ask Zeplik to check it for accessibility issues. A fragment is fine; you do not need the whole document. You do not have to open the tool explicitly; asking to scan markup for a11y issues is enough for Zeplik to run it.
Zeplik runs a real scan
It statically analyzes the markup against WCAG rules, not from a general impression, and treats the result as measured source data. The markup is handled as untrusted content, so nothing inside it can steer the answer.
Read the findings and fix
The answer lists each issue by rule, points at the offending elements, and explains the fix. Ask it to rewrite the markup, prioritize the findings, or re-scan after your changes, all in the same conversation.

Worked examples

Real prompts you can open in the composer with the text ready to review. Nothing sends on its own. Paste in your own markup.

Prompt: Scan this HTML for accessibility issues and tell me which WCAG rules it breaks: paste your markup here

Zeplik runs a real static scan over the markup and reports the measured findings by rule, so the answer is grounded in what the code actually contains, not a general impression.

Open this in Zeplik

Prompt: Check my page markup for missing alt text and unlabeled form fields

Ask about specific problems and the scan surfaces every offending element, so you get a concrete list to fix rather than vague advice.

Open this in Zeplik

Prompt: Does this button and background color combination meet WCAG AA contrast?

Give it the colors and it computes the actual contrast ratio and compares it against the AA threshold, so the pass or fail is measured, not eyeballed.

Open this in Zeplik

What it can and cannot do

The scan is precise about what it can measure and honest about its edges. Here is the plain posture.

Does itMissing alt text and labels

The scan finds images without alt text, form fields without labels, and other missing accessible names, and points at the exact elements.

Does itColor contrast against WCAG

It computes real contrast ratios for the colors you provide and checks them against the WCAG AA thresholds, so a pass or fail is measured.

Does itHeading order and structure

It flags skipped heading levels and structural issues that make a page hard to navigate with a screen reader.

Does itExplains the fix, not just the flaw

Each finding comes with what to change and why, so you can act on it rather than just log it.

Does notA live audit of a URL

The scan reads the HTML you paste, not a live rendered page. To check a page you have deployed, copy its markup, or use a full browser-based audit tool for the rendered result.

Does notGuaranteeing full WCAG compliance

Static analysis catches many issues but not everything a human tester would, like whether alt text is actually meaningful. Treat a clean scan as a strong start, not a certificate.

Grounded, and honest by default

It measures, not guesses. Findings come from a real static scan of your markup, tied to specific elements and rules, so you can act on each one.
It is honest about coverage. Static analysis cannot catch everything a human tester would, and Zeplik says so rather than implying a clean scan means full compliance.
Your markup cannot hijack the answer. The HTML is treated as untrusted content: Zeplik scans and reports it but never obeys instructions embedded inside it.

Frequently asked questions

Why not just ask a chatbot to review my HTML?: Because a language model reading markup cold will give an impression, and can miss or invent issues. This tool runs a real static scan against WCAG rules and reports measured findings, pointing at the actual elements, so the answer is grounded in what the code contains rather than a guess.
Which issues does it catch?: It covers common WCAG failures that static analysis can detect: missing alt text, unlabeled form fields and controls, low color contrast, skipped heading levels, and similar structural problems. Each finding is tied to a rule and the element that breaks it.
Does the scan cost credits?: The scan itself is free. Running it draws nothing from your balance; you only pay the normal model-token cost of the answer Zeplik writes around it, exactly like any other message, and the cost is shown before you send.
Can it check a live URL instead of pasted HTML?: The scan works on the HTML you paste, not a live rendered page. For a deployed page, copy its markup into the chat, or pair this with a full browser-based audit tool that evaluates the rendered result. Zeplik can also read a page's text if you give it the link.
Does a clean scan mean my page is fully accessible?: No. Static analysis catches many issues but not everything, and some things, like whether alt text is actually meaningful or the tab order makes sense, need a human or assistive-technology test. Treat a clean scan as a strong start, not a compliance certificate.
Which AI model does the scan?: The scan is a built-in capability that works the same regardless of model. The explanation around it is written by whichever of Zeplik's frontier chat models you pick, and you can switch models at any time, including mid-conversation.

Related skills

Go deeper on inclusive, well-built interfaces with these ready-to-run Zeplik skills.

More on Zeplik

Catch the issues before your users do

Scan your HTML for real WCAG issues and get the fix for each, grounded in a measured scan. Free to start.

Scan for accessibility issues Browse all tools