MCP server documentation

MCP server documentation

"monsido-mcp-server": {
 "type": "remote",
 "url": "https://api.monsido.com/mcp",
 "headers": {
   "Authorization": "Bearer <JWT token>"
 },
 "enabled": true
}

The server advertises support for the following: 

• Tools (with listChanged)

• Prompts (with listChanged)

• Resources

• Logging

• UI extension (`io.modelcontextprotocol/ui`)

Tools

In the Web Governance suite,13 tools are live. These are grouped by capability area as per the following:

Accessibility tools

The following sections provide information about the accessibility tools.

HTML scan

html_scan

• Title: HTML Scan
• Tags: accessibility, html, scan
• Read-only: yes
• Description: The system does a scan on the raw HTML and its page-processing-lib encoded_page for accessibility, readability, SEO, and data-protection findings. Use this tool when you already have processed page data. Use render_by_url_with_guideline for live rendered URL accessibility checks.

Input

Output

Usage notes

• Both html and encoded_page are required, and these must correspond to the same page.

• Omit accessibility to get readability, SEO, and data protection results without a WCAG scan.

Render by URL with guideline

render_by_url_with_guideline

• Title: Render By URL With Guideline

• Tags: accessibility, scan, url, wcag

• Read-only: yes

• Description: The system renders a live URL in a browser and checks accessibility against an optional WCAG guideline. Use this for public pages where the rendered DOM is needed. Returns rendered HTML, HTTP status, accessibility summary, and grouped errors.

Input

Output

This output includes rendered HTML, HTTP status, accessibility summary, and grouped errors (schema varies by guideline and the max_errors setting).

Usage notes

• Preferred over html_scans when you only have a URL and no pre-processed page data.

• JavaScript rendering is applied automatically. Use this for SPAs and dynamically-rendered pages.

Accessibility remediation source code

accessibility_remediation_source_code

• Title:

  Accessibility
  Remediation Source Code

• Tags: accessibility, remediation, source_code

• Read-only: no

• Description: If there are one or more HTML snippets and an accessibility check identifier, the system returns each fixed snippet with a plain-language description that explains what changed and the reason for the change.

Input

Output

This output returns a result array, one item per input snippet:

Usage notes

• Multiple snippets can be sent in a single call; results are returned in the same order.

• Always do a new scan after you apply fixes in order to confirm a clean resolution.

Spelling tools

The platform provides tree variants. Use the one that matches your available input. All of the variants require the Spell Check (qa_spellcheck) plan feature.

Spellling text snippet

spelling_text_snippet

• Title: Spelling Text Snippet

• Tags: check, content, spelling

• Read-only: yes

• Description: The system checks for spelling errors in a plain text snippet. The system returns misspelled words and the language that they were checked against.

Input

Output

This output returns a result array of spelling errors with misspelled words and language context.

Usage notes

Best for targeted checks, such as:

• A specific paragraph

• A flagged term from a policy violation

• A sentence that is under review

Spelling URL

spelling_url

• Title: Spelling URL

• Tags: check, content, spelling

• Read-only: yes

• Description: The system checks for spelling errors on a live URL. The crawler does a scan on the page, detects the language of each text block, and returns misspelled words.

Input

Output

This output returns a result array of spelling errors across the page.

Usage notes

• Use these for live site monitoring or pre-publish checks when you have a URL rather than raw HTML.

• Enable enable_javascript for dynamically-rendered content.

Spelling page ID

spelling_page_id

• Title: Spelling Page ID

• Tags: check, content, spelling

• Read-only: yes

• Description: The system checks for misspelled words on a stored Web Governance page using data from the last successful scan. This option is fast. The system does not do a live network request.

Input

Output

This output returns a result array of spelling errors from the last crawl.

Usage notes

• This is the fastest spelling option. The system returns stored results, it does not do a new crawl.

• Use resolve_url_to_ids to get domain_id and page_id from a URL.

Readability tools

Three variants. All variants return the same core output form.

Readability text snippet

readability_text_snippet

• Title: Readability Text Snippet

• Tags:content, readability, score

• Read-only: yes

• Description: The system calculates readability by analyzing raw text directly.

Input

Output

This output returns the following:

Usage notes

• Best for targeted analysis such as a specific paragraph, a policy-flagged passage, or a sentence flagged by html_scans.

• level == "Legal / Technical" is a strong signal that complexity is intentiona. This is useful for policy triage.

Readability URL

readability_url

• Title: Readability URL

• Tags: check, content, readability

• Read-only: yes

• Description: The system calculates the readability for a live public URL. Returns score, level, word count, analyzed text, and reason.

Input

Output

This output returns the following:

Readability page

readability_page

• Title: Readability Page

• Tags: page, readability, score

• Read-only: yes

• Description: The system retrieves readability for a page already stored in Web Governance. This option is fast. The system returns the results from the most recent successful scan, it does not do a new scan on the page.

Input

Output

Usage notes

• Use resolve_url_to_ids to get domain_id and page_id from a URL.

Policy tools

There are tree variants. All require the

Run policy snippet

run_policies_snippet

• Title: Run Policy (Snippet)

• Tags: check, content, governance, policies, snippet

• Read-only: yes

• Description: The system evaluates one governance policy against draft text or HTML before it does the crawl. Use this option for body-copy, keyword, regex, page_html, and page_text checks.

Input

Output

This output returns the following:

Run policy URL

run_policies_url

• Title: Run Policy (URL)

• Tags:check, content, governance, policies, url

• Read-only: yes

• Description: The system fetches a live URL and evaluates one governance policy against its fetched HTML, text, title, URL, and links. Use this option when freshness matters or the page has not been crawled in Web Governance.

Input

Output

This output has the same form as run_policies_snippet.

Run policy page ID

run_policies_page_id

• Title: Run Policy (Page ID)

• Tags: check, content, governance, policies

• Read-only: yes

• Description: The system dvaluates one governance policy against a page that has been previously crawled by Web Governance, using the content from the most recent successful crawl. This option supports a full stored-page evaluation that includes domain targeting.

Input

Output

This output has the same form as the run_policies_snippet.

Utility tool

Resolve URL to IDs

resolve_url_to_ids

• Title: Resolve URL to IDs

• Tags: domains, ids, pages, resolve

• Read-only: yes

• Description: Resolve a known URL to Web Governance domain_id / page_id values, needed by the _page_id and _page variants of spelling, readability, and policy tools.

Input

Output

The system groups the results by domain, each domain contains matching pages with their domain_id and page_id. If truncated=true, retry with domain_ids and/or a more specific URL.

Usage notes

• This is the entry point for any workflow that starts with a URL and needs to call a _page_id tool.

• Use partial=true when you have a partial URL or want to find all pages under a path.

Resources and prompts

The server currently exposes no resources and no prompts.

Example workflows

Accessibility check and fix from a URL

1. render_by_url_with_guideline(url, accessibility_guideline="WCAG22-AA")
  → Collect: accessibility errors with check_ids and source snippets
2. accessibility_remediation_source_code(check_id, source_codes)
  → Collect: fixed HTML snippets with descriptions
3. Re-run render_by_url_with_guideline to confirm clean resolution

Full content quality check on a live URL

1. render_by_url_with_guideline(url, accessibility_guideline="WCAG22-AA")
  → Accessibility errors
2. spelling_url(url)
  → Spelling errors
3. readability_url(url)
  → Readability score and flagged sentences
4. run_policies_url(url, policy_id) [repeat per policy]
  → Policy violations
5. accessibility_remediation_source_code for auto-fixable errors

Checking a draft before publish (no live URL yet)

1. html_scans(html, encoded_page, accessibility="WCAG22-AA")
  → Full scan: accessibility, SEO, readability, data protection
2. spelling_text_snippet(text)
  → Spelling errors in the draft body copy
3. run_policies_snippet(policy_id, content=html, content_type="html") [repeat per policy]
  → Policy violations against draft HTML
4. For any readability concerns: readability_text_snippet(text)
  → Targeted readability score on specific paragraphs

Bulk audit using stored Web Governance data

1. resolve_url_to_ids(url)
  → Get domain_id and page_id
2. spelling_page_id(domain_id, page_id)
  → Spelling errors from last crawl (fast, no network call)
3. readability_page(domain_id, page_id)
  → Readability score from last crawl
4. run_policies_page_id(domain_id, page_id, policy_id) [repeat per policy]
  → Policy violations from last crawl

spelling_url

Read-only: no

subject