window.c15t API

Call c15t before and after the tag loads

window.c15t is the only global the script tag uses. Before the tag runs, it is an array of calls. Push [method, ...args] entries onto it:

<script>
  window.c15t = window.c15t || [];
  c15t.push(['config', { legalLinks: { privacyPolicy: { href: '/privacy' } } }]);
  c15t.push(['on', 'ui', (surface) => console.log('surface:', surface)]);
</script>

When the tag runs, it replaces the array with the API object and replays the calls in order. Always push onto the existing array. window.c15t = [[...]] throws away calls that other scripts queued earlier, including the DevTools tag's.

c15t.push([...]) keeps working after the tag has loaded, and runs each call by the same rules as a queued one. So a snippet written as window.c15t = window.c15t || []; c15t.push([...]) works whether it runs before or after the tag. push returns the number of calls it received.

Which calls you can queue

A queued call that throws is reported with console.error, and the calls after it still run. A queue cannot hand back a return value, which is why reads are skipped.

How options merge

Options merge in this order, with later ones winning:

1. The tag's data-* attributes.
2. Each queued config call, in order.
3. Options passed to c15t.init(), when the tag has data-manual.

ui, ui.banner and ui.dialog merge key by key, so a config call that sets ui.theme keeps the legal links the tag's attributes set. Every other option replaces the earlier value.

config after the client has started logs a warning and does nothing.

Start c15t yourself

With data-manual on the tag, the bundle installs window.c15t and waits. c15t.js still requires a backend URL or hosted factory when init() runs. Use c15t.offline.js for browser-only policy resolution. Manifest and custom modes need the @c15t/browser ES module, a headless script with your own UI, or the IAB script for an IAB policy.

Call c15t.init() when your page is ready, for example once it knows the visitor's country:

<script>
  window.c15t = window.c15t || [];
  document.addEventListener('DOMContentLoaded', () => {
    c15t.push(['init', { overrides: { country: document.body.dataset.country } }]);
  });
</script>

init() reads the original tag's attributes even if the tag has since been removed, layers queued config calls and its own options on top, starts the client and returns it. A second call returns the same client. Until init(), queued actions and subscribe wait, and direct calls to methods that read or change consent throw call c15t.init() first.

After c15t.dispose(), a new c15t.init() starts a fresh client with the same attributes and queued config calls.

Read state

A permission is not a recorded choice

has('measurement') answers "may measurement code run now?" Under an opt-out policy it is true before the visitor has done anything. Use it, or snapshot.effectivePermissions, to gate code.

hasConsented() and snapshot.explicitChoice answer "what did the visitor decide?" They change only when the visitor accepts, rejects or saves. Use them to report or display the visitor's decision, never to gate a vendor. How consent works lists which to read for each task.

Answers can change once the policy resolves, so wait for c15t.ready() or a ready listener before you read them at page load.

Record a choice

Call these only in response to a visitor's click or key press, never at load.

acceptAll(), rejectAll(), save() and saveIAB() resolve to a result with an ok flag. The choice applies in the browser and the banner closes before the backend request finishes. When the request fails, ok is false, the choice stays in effect, and c15t retries the request later.

Show and hide the UI

In c15t.headless.js, showBanner, openDialog and closeDialog still change the surface c15t reports through the ui event, so your own banner can follow them.

Lifecycle

Properties

c15t.js reports pkg: '@c15t/browser' and mode: 'hosted' once started.

c15t.client.setOverrides({ country, region, language, gpc }) changes the location, language or GPC signal policy matching uses. The DevTools Location tab uses the same inputs.

Check it works

1. Open the browser console on a page with the tag and run await c15t.ready(). It resolves with a snapshot whose resolution.status is matched.
2. Run c15t.has('measurement') and c15t.hasConsented() before and after you click Reject All. Under an opt-in policy both start false; afterwards hasConsented() is true and has stays false.
3. Run c15t.push(['openDialog']). The dialog opens.