Free tool · No signup

Website audit MCP server

Give your AI assistant our free website checks. Add one URL to Claude Code, Cursor or VS Code, then ask it to audit a site, read a report or check an email domain.

Free · No account or API key · Read-only checks of public pages

Connect your assistant

One endpoint, streamable HTTP transport, JSON-RPC 2.0. Protocol versions 2025-06-18 and 2025-03-26. No account and no API key.

MCP endpoint

https://lastingweb.com/api/tools/mcp

Claude Code terminal

claude mcp add --transport http lastingweb https://lastingweb.com/api/tools/mcp

Add --scope project to share it with your team through .mcp.json. Check it with /mcp inside Claude Code.

Cursor .cursor/mcp.json

{
  "mcpServers": {
    "lastingweb": {
      "url": "https://lastingweb.com/api/tools/mcp"
    }
  }
}

Use ~/.cursor/mcp.json instead to enable it in every project.

VS Code .vscode/mcp.json

{
  "servers": {
    "lastingweb": {
      "type": "http",
      "url": "https://lastingweb.com/api/tools/mcp"
    }
  }
}

Then pick the tools in Copilot Chat's agent mode.

Test with curl

curl -s https://lastingweb.com/api/tools/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Other clients: add a remote MCP server of type HTTP (streamable) with the endpoint above.

Tools

Tool	Arguments	What it returns
`list_tools`	none	Every URL-based audit on this site: slug, name, description and the modules it runs.
`run_audit`	`url` (required), `tool` (default `growth-audit`), `wait_seconds` (default 120, max 170)	Score and grade, per-module scores, the top 10 issues with fixes and the report link — or a report id if the audit is still running.
`get_report`	`id`	The same summary for a finished report.
`check_email_domain`	`domain`	The email deliverability check: MX, SPF, DKIM, DMARC, MTA-STS, TLS-RPT and BIMI.

Try asking

“Run the LastingWeb growth audit on yourfirm.com and list the three most important fixes.”
“Check the security headers of staging.yourfirm.com with the http-headers-grader tool, then fix our nginx config.”
“After changing the template, run accessibility-checker on our public staging URL and fix what it finds.”
“Is yourfirm.com protected against email spoofing? Use check_email_domain.”

What we check and why

What it is

The Model Context Protocol (MCP) is an open standard that lets AI assistants call external tools. Our MCP server exposes the same audits that run on these pages, so an assistant such as Claude Code, Cursor or GitHub Copilot in VS Code can run a check while it works on your website, read the findings and suggest or make the fixes.

The server uses the streamable HTTP transport at a single endpoint, https://lastingweb.com/api/tools/mcp, speaks JSON-RPC 2.0 and supports protocol versions 2025-06-18 and 2025-03-26. It needs no account and no API key. The same fair-use rate limit applies as on the website.

Available tools

list_tools — lists every URL-based audit on this site with its modules and description.
run_audit — starts an audit for a URL (the growth audit by default, or any tool slug) and waits up to 170 seconds; returns the score, per-module scores, the top issues with fixes and a report link.
get_report — returns the summary of a finished report by its id.
check_email_domain — runs the email deliverability check (MX, SPF, DKIM, DMARC, MTA-STS, BIMI) for a domain.

Typical uses

Ask your assistant to run the accessibility checker on a staging URL after it changes a template and to fix what it finds. Let it check security headers after editing your server configuration, or the email records after a DNS change. Because results come back as structured data as well as a readable summary, an agent can compare scores before and after its own changes and stop when the issues are gone.

How to connect

In Claude Code, one command adds the server; in Cursor and VS Code you add a short JSON entry to the MCP configuration file of your project or user settings. The exact snippets are below, each with a copy button. After adding the server, ask your assistant something like: run the LastingWeb growth audit on yourfirm.com and list the three most important fixes. The assistant will ask for permission before it calls a tool for the first time.

Privacy and limits

Audits only read public pages, exactly as on the website, and reports are kept for 30 days behind an unguessable link. The tools never change anything on the audited site. Long audits that do not finish within the wait time return a report id that the assistant can fetch later with get_report.

Frequently asked questions

Do I need an API key?

No. The server is free and needs no account. Requests are rate-limited per IP address in the same way as audits started on the website.

Which AI assistants work with it?

Any client that supports remote MCP servers over streamable HTTP, including Claude Code, Claude Desktop (via a connector), Cursor and VS Code with GitHub Copilot. Configuration examples are on this page.

Can the assistant change my website through it?

No. All tools are read-only checks of public pages. Any change to your site is made by the assistant in your own code, under your control.

Is there a streaming (SSE) endpoint?

No. Each call is a single HTTP POST with a JSON response; a GET to the endpoint returns 405. Long audits return a report id to fetch later.

What data do you keep?

The same as for audits on the website: the report itself for 30 days. We do not store the prompts or conversations of your AI assistant.