Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@a11y-mcp-server Check accessibility of https://example.com"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
An MCP server that exposes axe-core accessibility auditing as tools — so any MCP-compatible agent (Claude Desktop, Claude Code, or any other MCP client) can audit HTML without a project-specific integration.
This grew out of agentic-design-tools's
a11y-review Claude Code skill: the skill is Claude Code-specific, this is the same auditing capability
as a standard protocol server any agent can call.
Tool | Description |
| Audit an inline HTML string |
| Audit a local |
| Fetch a URL and audit the raw (server-rendered) response HTML |
Each accepts an optional tags array to restrict which axe-core rule tags run, e.g. ["wcag2a", "wcag2aa"].
Scope note: audits run via jsdom, which has no real layout engine. That
reliably catches structural/semantic issues — missing alt text, unlabeled form inputs, missing
lang/<title>, empty link text, and similar — but rules that depend on rendered geometry or computed
color (color-contrast, focus-order checks) won't reliably fire. For full WCAG coverage, run axe-core in a
real browser (Playwright/Puppeteer) against the rendered page.
Security note: audit_html and audit_url execute any <script> in the input inside a sandboxed jsdom
window (no Node.js API access). Only audit HTML/URLs you trust.
Related MCP server: mcp-a11y-tools
npm install -g @aiuxmasters/a11y-mcp-serverAdd to your MCP client config (e.g. Claude Desktop's claude_desktop_config.json):
{
"mcpServers": {
"a11y": {
"command": "npx",
"args": ["-y", "@aiuxmasters/a11y-mcp-server"]
}
}
}import { auditHtml } from "@aiuxmasters/a11y-mcp-server";
const result = await auditHtml("<img src='logo.png'>");
console.log(result.violations); // [{ id: "image-alt", impact: "critical", ... }]npm install
npm test # unit tests + a real end-to-end MCP client/server round trip over InMemoryTransport
npm run buildMIT — see LICENSE.
Same checks as audit_html, reading the document from a local file path.
| Name | Required | Description | Default |
|---|---|---|---|
| path | Yes | Path to a local .html file | |
| tags | No | Restrict to specific axe-core tags, e.g. ["wcag2a", "wcag2aa"]. Omit to run all rules. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full burden. It only states 'same checks as audit_html' without disclosing behavioral traits such as side effects, permissions needed, or safety. Since it reads a local file, it is likely non-destructive, but this is not confirmed.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise with a single sentence. It is front-loaded and efficient, though it could provide more structured information. It earns its place but is slightly too terse.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the simple parameters and absence of an output schema, the description is minimally adequate. It relies heavily on knowledge of the sibling tool audit_html. The return format or behavior is not described, so completeness is moderate.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, and both parameters (path, tags) have clear descriptions. The tool description adds no additional meaning beyond what the schema already provides, so a baseline score of 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states that the tool performs the same accessibility checks as audit_html but reads from a local file path. This differentiates it from its sibling tool and specifies the exact action and resource.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Usage is implied by the description: use when you have a local HTML file. However, there is no explicit guidance on when not to use it or how it compares to audit_html for non-file contexts. The sibling tool is named but no exclusion criteria are provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Runs axe-core against an HTML string (via jsdom) and returns WCAG violations. Catches structural/semantic issues (missing alt text, unlabeled inputs, missing lang/title, empty link text, etc.) — layout-dependent rules like color-contrast are not reliable in jsdom.
| Name | Required | Description | Default |
|---|---|---|---|
| html | Yes | The HTML document to audit | |
| tags | No | Restrict to specific axe-core tags, e.g. ["wcag2a", "wcag2aa"]. Omit to run all rules. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description fully covers behavior: it uses jsdom, lists detectable issue categories, and honestly notes unreliable rules. No contradictions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two concise sentences with no unnecessary words. First sentence gives purpose, second adds important limitations. Well front-loaded.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (a11y auditing), the description covers purpose, method, limitations, and parameter usage adequately. No output schema but return type is implied. Complete for practical use.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description adds value by explaining the 'tags' parameter usage with examples like wcag2a. However, it doesn't specify that tags are optional or the default behavior if omitted.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it runs axe-core on an HTML string to find WCAG violations, and lists specific issues like missing alt text. This is a specific verb and resource, easily distinguishing from sibling tool audit_file which likely operates on files.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly states it catches structural/semantic issues but warns that layout-dependent rules like color-contrast are not reliable in jsdom. This provides clear guidance on when to use the tool and its limitations.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0audit_fileaudit_htmlScored across 2 tools
The two tools are clearly distinguished by input source (HTML string vs file path), but their core functionality is identical, which could cause minor confusion about which to use.
Both tools follow a consistent '_audit_' prefix pattern, making their purpose and variation (input type) predictable.
With only two tools, the server feels thin for a full accessibility auditing server, though it narrowly covers its stated input methods.
The server lacks features like rule configuration, output filtering, or detailed reporting, leaving significant gaps for practical usage beyond basic checks.
Deterministic axe-core accessibility scans (WCAG 2.1 AA, EN 301 549, PDF/UA) via your account.
Audit public webpages and supplied markup for HTML, CSS, SEO, JSON-LD, and link issues.
Scan a web page for accessibility, security, privacy, quality and SEO issues, with fixes.
Scan URLs for WCAG 2.1 violations, generate AI fixes, and produce VPAT 2.5 compliance reports.
