Record and Replay vs Spec Mocks
Short answer: Record-and-replay captures real traffic and plays it back, which is fast and limited to what happened; a spec mock is generated from the contract, so it covers the intended API including the edge cases that never occurred. The first is a recording; the second is a rendering.
What recording captures
The real requests and responses, which makes it a quick way to seed a mock from observed behavior. It cannot cover the failures and states that never happened.
What a spec mock covers
| Property | Recorded | Spec mock |
|---|---|---|
| Source | Observed traffic | Contract |
| Edge cases | Only what happened | Declared responses |
| Intended API | Indirectly | Directly |
| On change | Re-record | On save |
Why the edge cases matter
The bugs that hurt are in the 429, the empty list, and the malformed response that rarely happens. A spec mock includes them from the contract; a recording waits for them.
The honest takeaway
Use recording to seed from real behavior and the spec mock for the full surface. They are a starting point and a complete rendering.
Return to the Powerduck vs Mockoon comparison, or read the API mocking guide.