Amami MCP docs

Core analytics

Amami starts from the same product foundation as Umami: lightweight, privacy-minded website analytics for understanding how people use a site. The goal is not to remove the dashboard. The goal is to keep the familiar analytics model and add an AI-native layer on top.

What Amami tracks

Amami can help teams understand:

Websites and tracked domains.
Pageviews, visitors, sessions, bounce rate, and visit duration.
Top pages, entry pages, and viewed paths.
Referrers, channels, UTM campaigns, and traffic sources.
Countries, browsers, operating systems, and devices.
Custom events and conversion actions.
Realtime active visitors.
Date ranges and trend changes over time.

These are the same kinds of signals teams expect from Umami-style analytics. Amami makes them available to both the dashboard and AI agents.

Core dashboard areas

Use the dashboard to inspect:

Overview metrics: visitors, pageviews, visits, bounce rate, and visit duration.
Pages: top URLs, entry points, and content paths.
Referrers and campaigns: sources, UTM campaigns, and launch links.
Devices and environments: browser, operating system, device, and country.
Events: button clicks, form submits, install actions, activation steps, and checkout actions.
Realtime: active visitors, current pages, and live source changes.
Reports: funnels, journeys, retention, attribution, UTM, revenue, and performance signals when supported by your deployment.

Privacy-minded defaults

The inherited analytics model is designed for aggregate website analytics rather than personal identity tracking. That means Amami is a practical fit for teams that want useful product and marketing signals without turning basic traffic analysis into a heavy surveillance system.

What the dashboard is for

Use the dashboard when you want to inspect charts, share a view with teammates, manage websites, review settings, or visually explore reports.

Use MCP when you want an assistant to answer a specific question, connect a new project, compare trends, or recommend the next growth action from your data.

What Amami adds

Amami adds:

Browser-based MCP login or registration.
Local authorization and API-key setup for AI clients.
MCP tools for agents such as Codex, Claude, Cursor, Windsurf, Cline, Continue, Zed, VS Code, OpenCode, and custom scripts.
AI workflows that turn analytics signals into recommendations.
Write-enabled setup flows for creating websites and sending test events when explicitly allowed.

How to use dashboard and AI together

Use the dashboard when you need to visually inspect a chart or share a report.
Use MCP when you need an assistant to choose the right metrics, compare periods, explain movement, or recommend actions.
Use events and goals when you want the assistant to reason about conversion quality instead of only traffic volume.
Use realtime monitoring during launches and incidents.