MCP server for AI assistants

Connect Claude Code, the Claude app, Cursor or VS Code to the Statsy MCP server so your assistant answers questions from your real analytics numbers.

Statsy runs a Model Context Protocol server, so an AI assistant can read your traffic, revenue, funnels and visitors instead of guessing. You approve the connection in the dashboard, and the assistant gets exactly the role you have in the workspace.

The server URL is shown under Site settings › AI & CLI. For the hosted service it is https://statsy.co/mcp. It speaks streamable HTTP and is stateless, so there is nothing to keep running on your machine.

Connect your assistant

Claude Code

bash

claude mcp add --transport http statsy https://statsy.co/mcp

Then run /mcp, choose statsy and sign in. Your browser opens the dashboard to approve access.

Claude app

In Claude on the web or desktop, open Settings › Connectors, choose Add custom connector, paste the server URL, connect, and approve access in the window that opens.

Cursor

~/.cursor/mcp.json

{
  "mcpServers": {
    "statsy": {
      "url": "https://statsy.co/mcp"
    }
  }
}

A project-level .cursor/mcp.json works too. Then open Cursor Settings › MCP and sign in when prompted.

VS Code

.vscode/mcp.json

{
  "servers": {
    "statsy": {
      "type": "http",
      "url": "https://statsy.co/mcp"
    }
  }
}

Start the server from that file and sign in when VS Code asks.

Clients that cannot open a browser

Send an account token in a header instead of signing in.

bash

claude mcp add --transport http statsy https://statsy.co/mcp \
  --header "Authorization: Bearer sta_your_token"

What the assistant can do

Tool	Purpose
`list_sites`	Lists the workspace’s sites with id, public key, domain, time zone and currency. Start here.
`get_overview`	Headline metrics for a period, optionally compared with the previous one.
`get_timeseries`	One metric over time by hour, day, week or month.
`get_breakdown`	Top values of a dimension such as channel, page, country, campaign or keyword.
`get_attribution`	Revenue by channel under a chosen attribution model and lookback.
`list_funnels`	Saved funnels and their steps.
`get_funnel`	A step-by-step funnel report with drop-off and revenue.
`search_visitors`	Finds visitors by id, email, name, user id, country, city or path.
`get_journey`	One visitor’s full timeline of pageviews, goals, touches and payments.
`get_bots`	Bot and AI crawler requests by family, bot name or path.
`list_goals`	The goals configured for a site.
`create_goal`	Creates a pageview or event goal. Needs editor or owner.
`add_note`	Adds a chart annotation, such as a deploy or a launch. Needs editor or owner.

Three ready-made prompts also come with the server: a weekly review, “why did revenue change?” and “which channel to double down on?”.

Permissions and limits

Assistants get the role you have in this workspace. Viewers can read reports only; editors and owners can also create goals and notes. If your trial has ended, the read tools keep working and the write tools refuse. Revoke access any time by removing the connection in your assistant.

Connections are limited to 120 requests a minute. Site tokens (st_) are not accepted here — use an account token or sign in.

Tip. Good first questions: “Which channel brought the most revenue in the last 30 days, and how does that compare with the month before?” or “Why did conversions drop this week? Check channels, pages and countries.”

Frequently asked questions

Does Statsy have an MCP server?

Yes. Statsy runs a hosted Model Context Protocol server at https://statsy.co/mcp. Claude Code, the Claude app, Cursor, VS Code and any other MCP client that speaks streamable HTTP can connect to it and read traffic, revenue attribution, funnels, visitors and journeys for your sites.

What can an AI assistant do with Statsy over MCP?

It gets thirteen tools: list sites, overview, time series, breakdowns, revenue attribution under any of the five models, funnels and funnel reports, visitor search, one visitor’s journey, bot and AI crawler traffic, goals, and two writes for editors and owners: create a goal and add a chart note. Three prompts are included: a weekly review, why revenue changed, and which channel to double down on.

Is connecting an AI assistant to my analytics safe?

Access is approved by you in the dashboard, limited to your workspace, and carries exactly your role: viewers read, editors and owners can also create goals and notes. Nothing is granted by URL alone, connections are rate limited to 120 requests a minute, and you revoke access by removing the connection in the assistant.

Does the MCP server work without a browser sign-in?

Yes. Pass an account token (sta_) in an Authorization: Bearer header when you add the server. Site tokens are not accepted. This suits headless agents and servers that cannot open the OAuth window.

Which analytics questions work best over MCP?

Comparisons and causes: which channel brought the most paying customers last month versus the month before, where a funnel loses people, what one customer did before buying, or how much traffic AI crawlers send. The assistant can chain tools, so one question can span several reports.

Last updated 18 September 2026 · Markdown version