How does test design differ for an API endpoint versus a UI form for the same feature?

For a 'create user' feature, the API test design covers: valid payloads (each required field, optional fields), invalid payloads (missing required fields, wrong types, boundary values), duplicate email, concurrent requests, authentication failures, and correct HTTP status codes (201, 400, 409, 401).

The UI test design adds: form validation messages appear in the right place, required field indicators, client-side validation fires before the API call, the success state renders correctly, and error responses from the API surface as readable messages rather than stack traces.

The principle: push as much variation testing as possible to the API layer where it is faster and more reliable. Use the UI layer for the happy path, key error states, and tests that specifically verify the presentation — not for exhaustive input boundary testing.