Tracwell

Founders and small teams building web products need to know which traffic actually converts into signups and paying customers, not just page views. Today this means piecing together separate analytics tools, payment systems and manual attribution, often without a clean link between a visit and the revenue it eventually produced. Many existing analytics tools also do not expose data in a form that AI coding assistants like Claude, Cursor or Codex can query directly, forcing teams back into dashboards to get answers. Without session-linked revenue data, it is hard to say which source, campaign or landing page actually pays.

Tracwell is a product analytics tool that follows a visitor from first visit through signup and payment, and exposes the same data through a dashboard, an API and an MCP server for AI agents. Teams add a script tag to their site for automatic page view tracking, define custom events like signup_completed, and send purchase and refund data from their server with the visitor's session ID so revenue is linked back to the originating source and landing page. The dashboard offers acquisition, conversion, funnel, revenue and realtime views, plus an AI crawler/bot tracking feature that shows which AI assistants send traffic and which bots read pages. Users can also connect Claude Code, Cursor, Codex or VS Code via tracwell.app/mcp to ask questions in plain language, and with permission let the agent create goals and funnels (it cannot delete data or change settings). A cookieless Private mode is available for teams that want traffic numbers without storing browser identifiers, and past data can be imported from Umami, Databuddy, Vercel or Seline.

More traffic now arrives through ChatGPT and other AI assistants, and AI bots increasingly read websites, which the product addresses by tracking AI referral traffic and bot visits alongside normal analytics.