capture-api-response-test-fixture

You can capture real API responses by logging or saving raw provider outputs from generateText or streamText workflows, then storing these results under a __fixtures__ directory to stabilize response parsing tests.

Response parsing tests fail when expected outputs are guessed rather than captured from real provider API responses, creating brittle tests that mismatch actual provider behavior during generateText and streamText operations.

Yes, you can capture streamed raw chunks for streamText test fixtures by enabling raw chunk capture options during the provider response capture process and saving those chunks into the __fixtures__ directory.

The best way to store provider response fixtures is to save captured raw outputs into a __fixtures__ directory while aligning fixture naming conventions and test helper setup patterns with existing project tests.

Yes, you need to align fixture naming conventions and test helper setup patterns with your project's existing response tests so that captured provider outputs integrate seamlessly into the current parsing test workflows.

You should avoid using captured provider responses for test fixtures when your response parsing tests require dynamic or non-deterministic provider behavior, as saved raw outputs are static and meant to match specific generateText or streamText results.