Scenarios reference

Named server states. Useful for demos, e2e tests, and reproducing edge cases.

Define

await mockr({
  endpoints: [
    { url: '/api/users', data: [{ id: 1, name: 'Alice' }] },
  ],
  scenarios: {
    empty: (s) => { s.endpoint('/api/users').clear(); },
    crowded: (s) => {
      const users = s.endpoint('/api/users');
      users.insert({ name: 'Bob' });
      users.insert({ name: 'Carol' });
    },
    down: (s) => {
      s.endpoint('/api/users').handler = () => ({
        status: 503, body: { error: 'service down' },
      });
    },
  },
});

Each scenario is (server: MockrServer) => void | Promise<void>. It runs against a freshly-reset() server, so changes don't leak between switches.

Switch

await server.scenario('empty');     // programmatic
await server.scenario(null);        // back to baseline

POST /__mockr/scenario   { "name": "down" }
GET  /__mockr/scenario                          # current name or null

The HTTP control surface lets a Cypress / Playwright test flip scenarios per spec without restarting mockr.

Reset semantics

server.reset() (or switching to a different scenario) restores every endpoint's baseline via deep copy. dataFile endpoints reset to the latest file content, not the original — so editing the JSON between switches sticks.