Staats

無Cookie網頁分析，您的程式碼代理可讀取並記住。透過MCP提供流量、來源與漏斗數據，加上部署標記的前後比較、儲存的網站脈絡，以及主動異常偵測。遠端伺服器，無需安裝：一個金鑰涵蓋帳戶內所有網站。

造訪 staats.ai

文件

Getting Started

From zero to live answers in your agent's chat in under two minutes.

1. Create your account

Sign in at staats.ai with Google or GitHub. Every account starts on the free tier: 10,000 monthly events, 1 site, and all MCP tools enabled.

2. Connect your agent

In your account view under Connect your agent, grab the one-line connection snippet for your client (Claude Code, Cursor, Windsurf, Codex, or a raw URL). Your account key is embedded in the URL, so there are no headers or env variables to configure: paste the command and the connection is active.

One connection covers your entire account: every current project and any future ones you add. For manual configuration, see the MCP Server reference.

3. Add the script tag

Create a site in the dashboard (or ask your agent to run create_site), then drop the site's script tag into your page head:

<script defer data-site="YOUR_SITE_CODE" src="https://www.staats.ai/s.js"></script>

The script is ~1.5 KB compressed, sets no cookies, collects no personal identifiers, and requires no GDPR cookie banner. Pageviews begin streaming immediately, including client-side route changes in single-page apps.

Two things worth knowing up front: leave the snippet off your 404 page, and exclude your own visits so your browsing does not skew your own numbers.

4. Ask your first question

That's the entire setup. In your agent's chat, try:

  • "How's my site doing this week?"
  • "Where did yesterday's traffic spike come from?"
  • "Track clicks on the new signup button" (the agent adds a data-staats attribute; see Tracking)
  • "Annotate this deploy with the current commit and test back in 7 days" (see Agent Workflows)