Skip to main content
Glama

wcagc-mcp

WCAG-Compliance/wcagc-mcp MCP server

An MCP server that lets an AI assistant (Claude, ChatGPT, or any MCP-compatible client) run real, deterministic accessibility scans through wcagc — axe-core under the hood, not an LLM guess. Every scan result carries an explicit coverage disclaimer and never claims "compliant": automated testing finds only a portion of accessibility barriers, and this tool says so in every response.

This package is a thin, stateless adapter. It holds no database, no scan logic, and no secrets beyond the wcagc API base URL — it translates MCP tool calls into HTTP calls against the wcagc API and forwards the caller's own bearer. All authentication, entitlements, quotas, and scan orchestration live in the API; this code is safe to read end to end.

Two ways to run it

Local (stdio) — for Claude Desktop, Cursor, or any MCP client that spawns a local process:

npx @wcagc/mcp

Configure your MCP client with:

{
  "mcpServers": {
    "wcagc": {
      "command": "npx",
      "args": ["-y", "@wcagc/mcp"],
      "env": {
        "WCAGC_MCP_KEY": "<your mcp:scan API key>"
      }
    }
  }
}

Mint an mcp:scan key from your wcagc account under Settings → API keys — available on every plan, with a daily quota on Free/Starter and unlimited on Pro/Agency.

Hosted (Streamable HTTP + managed OAuth) — what Claude web/desktop/mobile connectors and ChatGPT use, since neither runs a local process for you. Add this remote MCP connector:

https://mcp.wcagc.com/mcp

The client discovers /.well-known/oauth-protected-resource/mcp, opens the wcagc login/consent flow, and binds the connection to one Organization. No key copy/paste is required. API-key bearer authentication remains supported for local stdio and CI.

ChatGPT availability depends on the ChatGPT plan and on whether the client permits action tools; scan_url creates a scan and is not a read-only operation. See wcagc.com/integrations/mcp.

Related MCP server: WCAG Color Contrast MCP Server

Tools

Tool

Plan

What it does

scan_url

all

Scan any public URL, or a registered site for full tracking (Pro+).

check_pdf

all

Run a PDF/UA-1 structure check on a public PDF.

get_scan · get_findings

all

Read a scan_url scan's status, severity counts, and findings — either kind, recorded or one-off.

list_sites

Pro+

List the account's registered sites.

scan_site

Pro+

Crawl and scan every reachable page of a registered site.

get_run · get_run_findings

Pro+

Read a full-site run from scan_site.

get_root_causes

Pro+

Group a run's repeated DOM patterns and return factual element/page blast radius.

run_journey

Pro+

Replay a saved multi-step journey and check each step.

get_trends

Pro+

Read a site's violation-count history over time.

One id, one poll tool: whatever scan_url did with a URL, get_scan and get_findings read it back. get_run and get_run_findings are only for full-site runs from scan_site.

Every scan-producing tool returns the coverage disclaimer in both the text content and the structured content. There is no score, grade, or conformance verdict — automated testing finds roughly 30–57% of accessibility issues, and the remainder needs manual review.

Configuration

Env var

Used by

Meaning

WCAGC_API_BASE_URL

both

The wcagc API to call. Defaults to https://api.wcagc.com; set it only when self-hosting.

WCAGC_MCP_KEY

stdio

Your mcp:scan API key.

PORT

hosted

Port to listen on (default 8080).

WCAGC_MCP_ALLOWED_HOSTS

hosted

Comma-separated Host-header allowlist (DNS-rebinding protection when bound to 0.0.0.0).

WCAGC_MCP_INTROSPECT_TTL_SECONDS

hosted

How long a verified bearer is cached before re-checking with the API (default 60).

WCAGC_MCP_OAUTH_ISSUER

hosted

Expected OAuth issuer (defaults to WCAGC_API_BASE_URL).

WCAGC_MCP_OAUTH_JWKS_URL

hosted

Authorization Server public JWKS URL.

WCAGC_MCP_URL

hosted

Canonical RFC 9728 protected-resource URL (defaults to https://mcp.wcagc.com/mcp).

WCAGC_MCP_JWKS_TTL_SECONDS

hosted

JWKS cache TTL; an unknown kid triggers an immediate refetch.

WCAGC_OPENAI_APPS_CHALLENGE

hosted

OpenAI Plugins Directory domain-verification token; keep it in the deployment secret store, never in source.

Development

Hosted HTTP requests to /mcp and discovery endpoints produce JSON logs with a generated request ID, HTTP status, MCP method, error code and duration. Tokens, headers, query strings, tool arguments and results are excluded. Health probes are not logged.

The Docker entrypoint preloads the New Relic ESM agent. Set NEW_RELIC_AGENT_ENABLED=true, NEW_RELIC_APP_NAME=wcagc-mcp and the deployment secret NEW_RELIC_LICENSE_KEY to enable APM and forwarding of these safe request logs. The license key must belong to the existing New Relic account; never commit it. The stdio entrypoint does not load the agent.

npm install
npm run dev        # hosted, watch mode
npm run start:stdio # stdio mode
npm run typecheck
npm run verify      # node:test against a local fixture API

License

MIT — see LICENSE.

Available Tools

16 tools
check_pdfCheck a PDF for PDF/UA-1 conformanceAInspect

Downloads a public PDF URL and runs a machine-verifiable veraPDF PDF/UA-1 check (document structure, tags, reading order — not a full WCAG audit). Free tier, counted against the caller's daily quota. Queues the check and returns immediately with an id; call get_pdf_check to poll.

Inputs: url (required). Returns: pdfCheck, standard, scoringGuidance, coverageDisclaimer

get_findingsGet a scan's findingsA
Read-onlyIdempotent
Inspect

The findings for a scan started by scan_url — rule id, severity, help URL and the element selector. Works for both kinds of scan_url scan. A one-off scan returns the top-5 sample; a scan recorded against a registered site returns its full findings. For a full-site run from scan_site, use get_run_findings instead.

Inputs: scanId (required). Returns: violations, recordedAgainstSite

get_fixesList tracked fixesA
Read-onlyIdempotent
Inspect

Lists tracked remediation items for a registered site, including status and the scope of any automated verification proof. Pro+ (REMEDIATION_TRACKING).

Inputs: siteHost (required). Returns: fixes, standard, scoringGuidance, coverageDisclaimer

get_fix_verificationGet fix verificationB
Read-onlyIdempotent
Inspect

Polls a targeted fix verification and reports only what was detected on the selected pages checked. Pro+.

Inputs: fixVerificationId (required). Returns: standard, verification, scoringGuidance, coverageDisclaimer

get_journey_runGet a journey run by idA
Read-onlyIdempotent
Inspect

Polls a run started by run_journey — per-step checkpoints and, once terminal, a failure reason if a step failed. Pro+.

Inputs: runId (required). Returns: run, standard, scoringGuidance, coverageDisclaimer

get_pdf_checkGet a PDF check by idA
Read-onlyIdempotent
Inspect

Polls a PDF check started by check_pdf — status, failed-rule and failed-check counts, and a failure reason if it did not complete.

Inputs: checkId (required). Returns: pdfCheck, standard, scoringGuidance, coverageDisclaimer

get_root_causesGet a scan run's root causesA
Read-onlyIdempotent
Inspect

Returns deterministic repeated DOM patterns and factual blast-radius counts for a full-site run. Grouping does not expand automated-test coverage and may not match the site's real component boundaries. Pro+.

Inputs: runId (required). Returns: rootCauses

get_runGet a full-site scan run by idA
Read-onlyIdempotent
Inspect

Polls a full-site run started by scan_site — page progress, severity counts and, once terminal, a failure reason. Pro+. For a single-page scan from scan_url, use get_scan instead.

Inputs: runId (required). Returns: run, standard, scoringGuidance, coverageDisclaimer

get_run_findingsGet a full-site scan run's findingsA
Read-onlyIdempotent
Inspect

Run-level, rule-deduplicated findings for a scan_site run — one entry per rule, with the WCAG success criteria it maps to. Pro+. For a single-page scan from scan_url, use get_findings instead.

Inputs: runId (required). Returns: violations

get_scanGet a scan by idA
Read-onlyIdempotent
Inspect

Polls a scan started by scan_url — either kind, recorded against a registered site or one-off; this tool finds it either way. Returns severity counts, a finding sample, the coverage disclaimer, and (once terminal) a failure reason if the scan did not complete. For a full-site run from scan_site, use get_run instead.

Inputs: scanId (required). Returns: scan, pollWith, standard, scoringGuidance, +2 more

list_sitesList registered sitesA
Read-onlyIdempotent
Inspect

Lists this organization's registered sites — their normalized hosts are what every other Pro+ tool takes as siteHost. Pro+ (requires sites:read + API_ACCESS).

Inputs: no named parameters. Returns: sites

run_journeyRun a saved user journeyA
Destructive
Inspect

Replays a saved multi-step journey (e.g. "add to cart -> checkout") against its registered site and checkpoints accessibility at each step. Pro+ (JOURNEYS; also AUTHENTICATED_SCANS when the journey logs in). Steps and any login credential always come from the journey's own saved configuration — never accepted here. Saved click or fill steps can submit forms or trigger external actions, so run only a journey the user has reviewed. Queues the run and returns immediately with a runId — call get_journey_run to poll.

Inputs: siteHost (required), journeyName (required). Returns: run, standard, scoringGuidance, coverageDisclaimer

scan_siteStart a full-site scanAInspect

Crawls and scans every reachable page of a registered site. Pro+ (SCAN_FULL_SITE). Queues the run and returns immediately with a runId — call get_run to poll. Call list_sites first if you do not already know the exact registered host.

Inputs: siteHost (required). Returns: run, standard, scoringGuidance, coverageDisclaimer

scan_urlScan a URLAInspect

Scans one http(s) URL for accessibility problems with axe-core — deterministic checks, not a language model's opinion. Pass the URL and nothing else. If it belongs to a site registered in the caller's account (and their plan allows), the scan is recorded against that site so it feeds history and trends; otherwise it runs as a one-off. Either way, poll it with get_scan. Returns immediately with an id — never a compliance score, because automated testing finds only a portion of accessibility barriers; see coverageDisclaimer in the result.

Inputs: url (required). Returns: scan, pollWith, standard, scoringGuidance, +2 more

verify_fixVerify a tracked fixAInspect

Queues a targeted automated re-check of representative pages for one tracked fix. This is not a full-site re-scan or a compliance guarantee. Pro+.

Inputs: remediationItemId (required). Returns: pollWith, standard, verification, scoringGuidance, +1 more

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.9/5.0

    Scored across 16 tools

    Disambiguation4/5

    Most tools are clearly scoped to distinct resources and actions: scan_url/scan_site, check_pdf, run_journey, and verify_fix each have their own dedicated polling tool. The many get_* tools could be confused at a glance, but descriptions explicitly cross-reference which poller to use for each job type.

    Naming Consistency5/5

    All tools follow a consistent lowercase snake_case verb_noun pattern (scan_url, list_sites, get_run_findings, verify_fix). Paired queue/poll tools are predictably named (check_pdf -> get_pdf_check, run_journey -> get_journey_run), making the set easy to navigate.

    Tool Count4/5

    16 tools is slightly above the ideal 3-15 range but each tool maps to a distinct stage in the scanning pipeline: initiating a job, polling it, retrieving findings, and listing supporting resources. The count feels justified by the breadth of scan types rather than redundant.

    Completeness4/5

    The surface covers the core lifecycle for URL scans, full-site runs, PDF checks, journeys, trend history, and fix verification. Minor gaps exist — no site registration/update/delete and no cancellation or re-run operations — but these are likely managed outside the MCP and don't block primary workflows.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      D
      maintenance
      Provides comprehensive access to WCAG 2.2 accessibility guidelines, including all 87 success criteria with full Understanding documentation, 400+ techniques, glossary terms, and ACT test rules from official W3C data.
      20
      86 npm
      15
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables accurate WCAG color contrast checking and accessibility analysis by calculating actual contrast ratios between color pairs, checking compliance levels, and analyzing luminance values for web accessibility.
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Provides AI agents with web accessibility analysis tools via MCP, enabling checks for alt text, heading hierarchy, color contrast, ARIA validation, and form accessibility.
      39 npm
      MIT