sitelens-mcp
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., "@sitelens-mcp export a severity-ordered fix checklist from my latest report"
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.
sitelens-mcp
Local companion MCP for the SiteLens Chrome extension.
Flow (no manual export):
Cursor starts this MCP over stdio (see config below).
MCP also listens on
127.0.0.1:19527(bridge).In Chrome, run SiteLens analysis → click Send to MCP on the floating panel.
The extension
POSTs the currentAnalysisReportto the bridge.Agent tools (
sitelens_get_latest_report, …) read the in-memory cache.
Everything stays on your machine. No cloud upload.
Install
git clone https://github.com/SiteLens-team/sitelens-mcp.git
cd sitelens-mcp
pnpm install
pnpm buildRequires Node.js ≥ 18.
Related MCP server: AgenticSEO MCP Server
Cursor mcp.json
Point at the built entry (use your absolute path):
{
"mcpServers": {
"sitelens": {
"command": "node",
"args": [
"/absolute/path/to/sitelens-mcp/dist/index.js"
]
}
}
}Optional port override:
{
"mcpServers": {
"sitelens": {
"command": "node",
"args": [
"/absolute/path/to/sitelens-mcp/dist/index.js"
],
"env": {
"SITELENS_MCP_PORT": "19527"
}
}
}
}Pairing
Step | Who | What |
1 | Cursor | Starts |
2 | You | Analyze a page with the SiteLens extension |
3 | You | Click Send to MCP on the panel toolbar |
4 | Extension |
|
5 | Agent | Calls |
Health probe:
curl -s http://127.0.0.1:19527/healthTools
Tool | Purpose |
| Bridge up?, port, cached report count |
| Full latest |
| Summaries only (last ~20) |
| Severity-ordered fix checklist (no AI) |
Environment
Variable | Default | Meaning |
|
| Localhost bridge port |
The Chrome extension defaults to the same port. Optional override in extension storage key mcpPort.
Privacy
Bridge binds only
127.0.0.1.Logs never include full report bodies (url + timestamp only).
Reports live in MCP process memory (ring buffer of 20); they vanish when the MCP process exits.
License
MIT
Available Tools
4 toolssitelens_export_checklistA
Build a severity-ordered actionable checklist from the latest (or url-matched) report: SEO warn/fail, image suggestions, a11y, security. No AI.
Inputs: url.
sitelens_get_latest_reportA
Return the latest AnalysisReport JSON pushed by the SiteLens extension. Optional url substring filter.
Inputs: url.
sitelens_list_reportsA
List cached SiteLens reports as summaries only (url, analyzedAt, score, module issue counts). Newest first.
Inputs: no named parameters.
sitelens_pingA
Health check: whether the local SiteLens bridge is up, which port it uses, and how many reports are cached.
Inputs: no named parameters.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool targets a distinct action: health check (ping), retrieve the newest full report, enumerate cached report summaries, and derive a checklist. The only near-overlap (get_latest_report vs list_reports) is clearly separated by 'full AnalysisReport JSON' versus 'summaries only, newest first.' An agent can select correctly from names alone.
All four tools share the sitelens_ prefix and a verb_noun pattern (get_latest_report, list_reports, export_checklist), which is predictable. The lone deviation is sitelens_ping, a bare verb with no noun, but that is idiomatic for a health check.
Four tools is on the lean side but each covers a genuinely separate need (health, fetch, enumerate, transform). Nothing appears redundant, though the surface is thin enough that a couple of convenience tools could reasonably have been added.
The domain is a read-only bridge over cached reports, and the surface covers discovery (list), retrieval (get latest, with url filtering), health, and one derived output (checklist). Minor gap: there is no way to fetch an older or specific report by id/url beyond 'latest matching,' and no cache-clearing operation.
Related MCP Connectors
Accessibility and WCAG data for your own websites: fix lists, live checks, and fix validation.
Scan URLs or HTML for WCAG 2.2 violations. 75-rule manifest, weighted score, shareable reports.
Automated accessibility check (axe-core, Lighthouse) of one public web page; fetch saved results.
Website health, SEO, security-header, SSL and broken-link audits for AI agents.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceEnables auditing of websites for performance, SEO, accessibility, security, and mobile readiness, with tools to validate URLs, run page audits, save results, and retrieve reports.1-
- AlicenseNot gradedqualityCmaintenanceEnables coding agents to audit public or localhost pages, gather structured SEO evidence, and retrieve compact reports, finding details, and next actions for triage.MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI assistants to analyze public websites, follow progress, retrieve shareable reports, compare multiple URLs, list checks and recent activity, and surface fixes ranked by impact.Apache 2.0

a11yscopeofficial
FlicenseNot gradedqualityBmaintenanceWCAG 2.2 AA accessibility checks for a URL or for HTML the agent just wrote, run locally in headless Chrome. 31 checks with CSS selectors, snippets and fix guidance; the same engine as the A11yScope Chrome extension.1-