Tracium

Docs

How it works

Two signals. A script for people who load the page, and an edge proxy for everything else, including crawlers that never run JavaScript.

1. Add a project

Create a project with the domain you want to watch. Tracium issues a certificate for that domain and for www.

2. Point DNS at the proxy

Enable the proxy on the project. Tracium looks up where the domain points and remembers that address. Then add a CNAME. Traffic hits the proxy first, gets recorded, and is forwarded there.

Type
CNAME
Name
www, and the apex if your DNS allows it
Target
proxy.tracium.dev

Once the hostname is active, the certificate is active too. Requests are logged before they reach your server, so a bot is recorded even when it does not execute a script.

3. Add the script

The proxy serves the tracker on your own domain. One tag records page views, including client-side navigations.

<script src="/tracium/track.js" defer></script>

Custom events use the same visitor and session.

telemetry.trackEvent("signup", { plan: "pro" })

4. Read the dashboard

Visitors

A person is new when their first session falls in the range you selected, and returning when a later session does. Someone who arrives and comes back inside the same range counts as both.

Crawlers

Each request is named from its user agent. GPTBot stays GPTBot. Unknown agents are learned instead of being filed under a generic label.

Activity

The log badges the first page of a first visit as new, and a page view as returning only when that person has an earlier session.