mcpField
M
marlo-mcp

marlo-mcp

Local
github.com/Kumar2421Business & Marketing7Updated today

About marlo-mcp

marlo-mcp is a read-only MCP server for running Marlo SEO and technical audits on a single web page. It is intended for users of Claude, Cursor, and other AI assistants, and returns prioritized findings with evidence and suggested fixes.

Key features

  1. 01Audit one web page
  2. 02Return top findings with evidence
  3. 03Explain impact and suggested fixes
  4. 04Read-only page fetching

Use cases

  1. 01Review SEO before publishing a page
  2. 02Identify technical issues on a site page
  3. 03Prioritize fixes from an audit

MCP support

The server exposes the `audit_site` tool to MCP clients. It fetches the requested page and returns audit findings without modifying it.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"kumar2421-marlo": {
"command": "npx",
"args": [
"-y",
"marlo-mcp"
],
"env": {
"MARLO_BASE_URL": "<MARLO_BASE_URL>",
"MARLO_ACCESS_PASSWORD": "<MARLO_ACCESS_PASSWORD>"
}
}
}
}
Run commandnpx -y marlo-mcp
Environment variables to set
MARLO_BASE_URLBase URL of your own self-hosted Marlo. Leave empty to use the hosted app at https://app.marlo.mlforge.in.
MARLO_ACCESS_PASSWORDSecretThe MARLO_ACCESS_PASSWORD of your self-hosted Marlo, if you set one. Only sent to MARLO_BASE_URL.

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmmarlo-mcp0.1.0

Tools

1

Saved tool list · updated 1d ago

ParameterTypeRequiredDescription
urlstringYesThe page to audit, for example https://example.com or example.com

Details

From the project README

marlo-mcp

An MCP server that lets Claude, Cursor and other AI assistants audit a web page for SEO and technical issues with Marlo, the open-source AI CMO.

Ask your assistant things like:

  • "Audit https://mysite.com and tell me what to fix first."
  • "Check this page before I ship it."

It has one tool, audit_site. It is read-only: it fetches the page and changes nothing. It checks one page, and returns the top findings, each with evidence, why it matters and a concrete fix.

Install

Add it to your MCP client. For Claude Desktop, Cursor, Windsurf and similar clients:

{
  "mcpServers": {
    "marlo": {
      "command": "npx",
      "args": ["-y", "marlo-mcp"]
    }
  }
}

For Claude Code: claude mcp add marlo -- npx -y marlo-mcp

Requires Node.js 20 or newer.

Use your own Marlo server

By default the tool calls the hosted Marlo at https://app.marlo.mlforge.in, which rate limits public audits. To use your own self-hosted Marlo instead:

VariablePurpose
MARLO_BASE_URLYour server, for example http://localhost:3000
MARLO_ACCESS_PASSWORDYour server's MARLO_ACCESS_PASSWORD, if you set one. It is only ever sent to MARLO_BASE_URL, never to the hosted app.
{
  "mcpServers": {
    "marlo": {
      "command": "npx",
      "args": ["-y", "marlo-mcp"],
      "env": { "MARLO_BASE_URL": "http://localhost:3000" }
    }
  }
}

What it sends

Only the page address you ask about, to Marlo's public audit endpoint. No API keys, files or conversation text leave your machine. Private and local addresses are refused by the server.

Limits

  • One page per call, and the top three findings. The full product tracks findings over time, checks AI readiness (GEO), reads Search Console and opens draft fix PRs on GitHub.
  • The hosted endpoint is rate limited. If you hit the limit, the tool says so and when to retry.

Development

npm install
npm test
node bin/marlo-mcp.mjs   # speaks MCP over stdio

License: AGPL-3.0, same as Marlo.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.