@xenosystem/licence

The main-process entitlement check every XENO desktop app runs. It answers the door's question for one product (XENO AUTH - AUTH GATE DELTA.md §9):

| result | meaning | the door shows | |---|---|---| | licensed | allowed, verified or within 14-day grace | the app | | unlicensed | the server was reached and said no (reason: unauthenticated, plan_required, …) | sign-in, or "needs a plan" | | expired-offline | grace ran out and the server is still unreachable | "couldn't confirm your plan" | | update-required | this build is below the supported floor — asked even when signed out | "update to continue" |

Fail open on a network error, closed on an explicit refusal. A timeout or a 5xx never revokes anyone within grace; a 401/403 does, immediately.

Use

import { startLicence, xenoClientHeaders } from '@xenosystem/licence'

const licence = startLicence({
  product: 'canvas',
  version: app.getVersion(),
  getToken: () => account.getAccessToken(),   // null when signed out
  readCache: () => store.get('licence'),
  writeCache: (l) => store.set('licence', l),  // persist atomically
  onChange: (l) => pushToDoor(l),
})
// licence.refresh() after sign-in / sign-out; licence.stop() on quit.

🔴 Send xenoClientHeaders(product, version) on EVERY request to the platform, including your OIDC token requests. The server refuses a product's own sign-in client below its floor unless this header says the build is supported. A product that sends it only on the licence check locks itself out of sign-in the day a floor is set.

The default origin is https://xenostudio.ai. Not api.xenostudio.ai — that is the inference gateway, which does not serve these routes.

What it does not do

It stops casual copying. It does not stop a patched binary, and no client-side check ever has. The durable protection is server-side: sign-in, sync, the agent and inference are refused to builds below the floor.

Contract: xeno-platform/docs/LICENCE-ENFORCEMENT.md.