Solutions / MCP Server

TofuBofu inside your AI assistant

Run your AI visibility scan without leaving Claude

TofuBofu publishes a remote MCP server. Paste one URL into your AI assistant and it gains the ability to run a free AI visibility scan for any B2B company, right inside the conversation: no signup form first, no API key, nothing to install. The chat window your buyers use to research vendors becomes the place you check whether they can find you.

Free to connect, includes the free monthly scan ChatGPT Claude Perplexity Gemini Google AI Overviews Bing Copilot
51%
of B2B buyers now begin vendor research on an AI chatbot, up from 29% a year earlier (G2, 2026). The MCP server puts the visibility check inside that same chat window.

What the TofuBofu MCP server is, in plain words

MCP, the Model Context Protocol, is an open standard that lets AI assistants like Claude and ChatGPT call external tools mid-conversation. TofuBofu's server lives at https://tofubofu.com/mcp and exposes two tools. scan_ai_visibility starts a free scan for a domain: give it a company website and your work email, and it runs real buying-intent questions across the AI engines and checks who gets named. get_visibility_report reads the finished results a minute or two later: the score, the per-query findings, and a link to the full live report.

It is the same free scan the website runs, under the same contract: one free scan per email per month, the finished report is emailed to you, and the live report link works in any browser. There is no API key and no account needed up front. Your assistant simply asks the two questions the scan needs, the domain and your email, and hands back the results.

How to connect it to Claude

On claude.ai or the Claude desktop app: open Settings, go to Connectors, click Add custom connector, name it TofuBofu, and paste https://tofubofu.com/mcp as the URL. Custom connectors require a paid Claude plan. In Claude Code, one terminal command does it: claude mcp add --transport http tofubofu https://tofubofu.com/mcp.

Then just talk to it. In a new chat, try: "Use TofuBofu to check AI visibility for acme.com, my email is you@company.com." Claude calls the scan tool, waits for the engines to answer, and summarizes where the company is named, where it is absent, and which competitors AI recommends instead. Ask follow-up questions and Claude reads the report data to answer them.

ChatGPT, Cursor, and every other MCP client

The server speaks streamable HTTP, the current MCP transport standard, so it works in any client that supports remote MCP servers: ChatGPT's connector settings, Cursor, Windsurf, and the growing list of MCP-capable tools. The connection details are always the same, one URL and no credentials.

Developers who want to inspect the server directly can use the MCP Inspector (npx @modelcontextprotocol/inspector) and point it at the endpoint, or send JSON-RPC requests to it with plain curl. Opening the URL in a normal browser shows a friendly setup page rather than the protocol stream, so it is also safe to share the link with a colleague.

What you get

Everyone (free)

Both MCP tools work for anyone with the URL: a free scan per email per month across all 6 AI engines, with the report emailed and viewable live. No TofuBofu account is required to run the first scan.

Fix and Dominate

The scan is the front door. Tracking month over month, sentiment, the guided fixes, and the drafted content live in the full app, where the report created by your MCP scan is waiting when you sign in.

Find TofuBofu across the MCP ecosystem

Official MCP Registry Published and active as com.tofubofu/ai-visibility. The canonical registry MCP clients read from. mcp.so Listed in the mcp.so remote-server directory.

See where you stand first

Run a free scan and see which of these gaps you have today.

Get your free audit

Frequently asked questions

What is MCP?

The Model Context Protocol is an open standard, originally introduced by Anthropic, that lets AI assistants securely call external tools and data sources during a conversation. A remote MCP server is a URL the assistant connects to; once connected, the assistant can use the tools that server exposes, in TofuBofu's case running and reading AI visibility scans.

Is the TofuBofu MCP server free?

Yes. Connecting costs nothing and the scan it runs is the same free scan offered on tofubofu.com: one per email per month, all six AI engines, report emailed to you with a live link. Paid TofuBofu plans add weekly scans, more queries, sentiment, and drafted content, but none of that is required to use the MCP server.

Do I need an API key or a TofuBofu account first?

No. The server needs only two things, passed conversationally: the domain you want scanned and your work email. The email is where the finished report is sent and is how the free-scan limit is applied. You can create an account afterwards to claim the report and track it monthly.

Which AI assistants can use it?

Any client that supports remote MCP servers over streamable HTTP. Today that includes Claude (claude.ai custom connectors on paid plans, Claude Desktop, and Claude Code), ChatGPT's connectors, Cursor, Windsurf, and most other MCP-capable tools. The URL is the same everywhere: https://tofubofu.com/mcp.

What data does it collect?

Exactly what the scan needs and nothing else: the domain you asked to scan and the email you provided for delivery. There are no tracking cookies in the MCP flow and no other personal data is requested. The resulting report is the same one you would get from the website.

Why does it ask for my email?

Three reasons. The finished report is emailed to you so you do not lose it when the chat ends. The free-scan allowance, one per month, is applied per email. And if you later create a TofuBofu account with that email, the report is already attached to it.

I connected it. What do I actually say in the chat?

Ask naturally. "Run an AI visibility scan for acme.com, my email is you@company.com" starts the scan; the assistant will tell you it takes a minute or two. Then "get the visibility report" reads the results, and from there you can ask things like "which queries am I losing?" or "who does AI recommend instead of us?" and the assistant answers from the report data.

Sources and further reading

Related solutions

AI Visibility Scan → Comparison Pages → Blog Content →