Monitor reference
This page lists what Monitor checks, how it scores results, and its limits. For an introduction, see Monitor.
Site-wide checks
Site-wide checks run once per account against your website. The dashboard shows them in the AI readiness panel, grouped by area.
| Check | Area | Dashboard name | What it verifies |
|---|---|---|---|
account_robots_check | Reach | Crawl rules | robots.txt is published. |
ai_bot_access_check | Reach | AI crawler access | Which known AI bots robots.txt allows or blocks. |
account_llms_txt_check | Understand | Content guide for AI | llms.txt is published. |
account_mcp_check | Act | Tools for AI agents | A Model Context Protocol (MCP) manifest (mcp.json or an MCP server card) is published under /.well-known. |
account_webmcp_check | Act | In-browser agent actions | A WebMCP discovery surface is published. |
account_skill_md_check | Act | Agent instructions | SKILL.md or an agent skills index is published. |
Each site-wide check shows one of these states:
| State | Meaning |
|---|---|
| In place | The signal or interface is present. |
| Opportunity | It is missing, and adding it could help, depending on your goals. |
| Your policy | It is missing on purpose, for example AI crawlers blocked in robots.txt. |
| Couldn't verify | The check couldn't run, for example because the site's bot protection refused the request. |
| Not checked yet | There is no result yet. |
Page checks
Page checks run on every monitored URL. The dashboard shows them in the Bot readiness panel, grouped by category.
| Check | Category | What it verifies |
|---|---|---|
structured_data_check | Markup | JSON-LD is present and syntactically valid (it passes the Shapes Constraint Language (SHACL) constraints). |
status_check | Delivery | The page responds with HTTP 2xx or 3xx. |
performance_check | Delivery | Time to first byte (warning above 800 ms, fail above 1800 ms) and total response time (warning above 2500 ms, fail above 4000 ms). |
resource_404_check | Delivery | No sub-resources (scripts, images, styles) fail to load. |
title_check | SEO fundamentals | A <title> is present. |
canonical_check | SEO fundamentals | A canonical link is present. |
page_meta_check | SEO fundamentals | A meta description is present. |
wordlift_bootstrap_check | WordLift | The WordLift script is installed. |
Each page check returns OK, WARN or FAIL. Monitor also evaluates the expectations of every segment the page matches.
Segments
A segment matches a URL if any of its glob patterns or static URLs matches it. A URL can belong to several segments, and Monitor evaluates it against the expectations of each one.
| Matcher | Syntax |
|---|---|
| Glob pattern | * matches any characters and ? matches a single character, for example https://www.example.com/blog/*. |
| Static URL | The exact URL. |
Expectations
| Kind | Meaning |
|---|---|
| Type present | The entity type must appear on the page. |
| Type absent | The entity type must not appear on the page. |
| Property present | The property must be set on the given type. |
| Property absent | The property must not be set on the given type. |
Use the 2 absent kinds when a markup pattern is wrong for a kind of page. Property rules take a type and a property name. Nested properties use dot notation, for example publisher.name on a BlogPosting.
| Severity | Effect |
|---|---|
| Required | Affects the page status and counts toward the semantic data quality score. |
| Recommended | Affects the page status but does not count toward the score. |
Scores
AI readiness has no score. The 2 page-level panels are scored as follows:
| Score | Calculated as |
|---|---|
| Bot readiness | The share of page checks that pass, across all URLs. |
| Semantic data quality | The share of required expectation evaluations that pass, overall and per segment. Recommended rules affect status but not the score. |
The dashboard colors both scores by the same bands: green from 80%, amber from 50% and red below 50%.
Defaults and limits
| Setting | Value |
|---|---|
| Run frequency | Every 24 hours per account |
| Result history | 90 days |
| Segments per account | 25 |
| Glob patterns per segment | 20 |
| Static URLs per segment | 50 |
| Expectations per segment | 20 |
To change the run frequency or history for your account, contact your WordLift account team.
API
To read the latest results for a URL from your own code, use the Monitor API. See the Monitoring API Guide.
Current limitations
- You can't add pages from the dashboard. WordLift sets up the monitored pages for your account.
- Results are not instant. You can't start a run manually. Changes show up after the next scheduled run.
- Monitor reports issues but doesn't fix them. It shows where structured data falls short. Fixing it is a separate step.
- AI readiness is not yet available through the MCP Server. Bot readiness and semantic data quality are (see Ask your AI assistant about Monitor).
- You write expectations by hand. Monitor doesn't propose rules from your site's content.