@stagewright-ai/browser

The Stagewright browser SDK. It tells Stagewright which features people actually use, so walkthroughs and help content follow real usage, and it shows in-app guidance (nudges) that's checked against your product on every PR.

npm install @stagewright-ai/browser

import { stagewright } from "@stagewright-ai/browser";

stagewright.init({ key: "pk_sw_…" });
stagewright.identify(user.id, { role: user.role });
stagewright.group(account.id, { plan: account.plan });

Your publishable key (pk_sw_…) is under Insights → Usage data in Stagewright. It's safe to ship in client code.

What it captures

• Route changes, and clicks on elements your feature map knows (by test id, role and accessible name).
• Views of Stagewright content.
• Events you send with stagewright.track(event, properties).

No keystrokes, no form values, no session replay. Ids are hashed by the API before storage, and properties are allow-listed. If PostHog or Segment is already connected to Stagewright, you only need identify and group.

Options

stagewright.init({
  key: "pk_sw_…",
  autocapture: true,          // route changes and known clicks
  guidance: true,             // in-app nudges
  quietRoutes: ["/checkout"], // never show guidance here
  viewerToken,                // signed by your backend; ties views to a person
});

Other calls: page(), reset() (on sign-out), flush(), shutdown().

Script tag

<script src="https://api.stagewright.ai/sdk/v1/stagewright.js" data-key="pk_sw_…" defer></script>
<script>stagewright.identify(USER_ID, { role: ROLE })</script>

data-api-url, data-autocapture="false" and data-guidance="false" are optional. Calls queued on window.stagewright = [["identify", id, traits]] before the script loads are replayed.

License

MIT