AI agents

Your analytics, readable by an agent

An MCP server and a REST API, so Claude, Cursor or ChatGPT can answer questions about your traffic directly.

MCP server

Native tools for Claude, Cursor and any MCP client. JSON-RPC over POST.

REST API

Predictable HTTP endpoints, documented by an OpenAPI 3.1 document.

llms.txt

A machine-readable description of the product, including what it does not do.

Connect it in a minute

Make a read-only key in the dashboard, then paste this into your MCP client’s config. The key is all the authentication the server needs.

{
  "mcpServers": {
    "asuito": {
      "url": "https://asuito.com/api/mcp",
      "headers": { "Authorization": "Bearer spx_..." }
    }
  }
}

Or call the REST API directly:

curl -H "Authorization: Bearer spx_..." \
  "https://asuito.com/api/v1/stats/example.com?range=7d"

12 tools, all read-only

An agent can read everything and change nothing. That is deliberate: the failure mode of a confused agent with write access is somebody’s data, and there is no question worth answering that needs it.

list_sites

Every site on the account, with its domain and id.

get_traffic

Visitors, sessions and pageviews over a window.

get_breakdown

Any dimension ranked — path, referrer, country, browser, device.

get_sources

Where traffic came from, with AI referrals as their own channel.

get_top_pages

The pages doing the work.

get_revenue

Money attributed to sources, when a payment provider is connected.

get_live

Who is on the site right now.

get_badge

The public visitor figures behind the badge.

growth_report

Traffic, sources, pages and revenue in one call — the usual question, answered once.

list_keywords

Tracked search terms and their current positions.

list_mentions

Public brand mentions, filterable by urgency and category.

get_mention_summary

Counts by category and sentiment over a window.

API documentation OpenAPI document