network-mock-server

To mock API responses in Playwright, intercept network requests using page.route and the CDP Fetch domain, capture a baseline of real responses, and replay them deterministically to ensure stable, offline-capable test runs.

Yes, you can simulate network latency and errors per endpoint by intercepting API requests with Playwright, modifying the mocked response fields, and injecting custom delays or error status codes across page navigations.

Yes, intercepting network requests for deterministic mocking requires a connected Playwright MCP server and a Chromium-based browser for CDP access to intercept API requests made by target pages.

API mocks persist across page navigations by leveraging Playwright's page.route and the optional CDP Fetch domain to continuously intercept and replay captured baseline responses during automated test runs.

The best way to record baseline API responses is to capture real API traffic during an initial run to establish a stable reference, then replay and modify those responses deterministically in subsequent automated tests.