title: ClickStream Docs description: First-party analytics, identity resolution, and bot classification for modern web applications.

ClickStream

First-party analytics, cross-device identity, and real-time bot classification — built for operators who want to own their visitor data end-to-end and serve intelligent responses back to their users in under 50 ms.

This site is the developer reference for installing ClickStream, instrumenting your application, and consuming the real-time Signals API from your own code.

Who this is for

• Developers and technical founders installing ClickStream for the first time — start with the install guide.
• Application engineers integrating real-time visitor context into page logic — see the Signals API reference.
• Platform operators provisioning a first-party tracking domain — see the DNS setup walkthrough.
• Data + platform teams integrating with the event ingestion layer — see the event schema reference.

How ClickStream is organized

• collector-worker ingests events from opted-in visitors over first-party HTTPS.
• identity-graph resolves hashed identity signals into cross-device person records.
• dashboard (einstein.clickstream.com) exposes the operator surface: Intelligence, Traffic Quality, Visitors, Forms, Billing.
• @clickstream/sdk (this repo) is the client-side tracker.
• @clickstream/signals (this repo) is the client-side library for consuming real-time scores from your page code.

Conventions used in this documentation

• Commands in shell blocks assume a macOS or Linux environment and pnpm as the package manager.
• API tokens in code samples use the prefix cs_test_; production keys use cs_live_.
• Time values are in milliseconds unless otherwise noted.
• Score values are on a 0–100 scale unless otherwise noted.

Status

The documentation site is being filled in incrementally. Pages marked stub contain scaffolding only; content lands in subsequent waves as part of the market-readiness plan.