What a Collection is
A Collection is a saved, ordered sequence of requests you can run together — built from requests you've captured, chained into a repeatable workflow rather than replayed one at a time.
Passing values between requests
Pull a value out of one response — a token from a login response, an ID from a list endpoint — with a variable extraction, then reference it in every request after as {{authToken}}. Mark a variable a secret and it's masked in the UI and never written to disk in plain text, the right default for anything that looks like a credential.
Assertions
Attach assertions to a request in a Collection and Hollowport checks them automatically every time the Collection runs — six kinds, deliberately no more, covering the deterministic checks a debugging session actually reaches for rather than a general assertion language:
- Status — the response status is, isn't, is one of, or falls within a range of codes.
- Duration — the response came back under a maximum time.
- Size — the response body stayed under a maximum byte size.
- Header — a named header exists, equals, or contains a value.
- Body — the response body contains or doesn't contain a plain-text string.
- JSON path — a value at a given path in the response body exists, equals, or has a given count.
A regression that would otherwise slip by quietly — a field silently disappearing from a response, a status code drifting from 200 to 204, a response that gets slower over time — gets flagged instead.
Running a Collection
Run the whole sequence from the Collections tab. Each step's result — pass, fail, and which assertion failed if any — is shown per request, not just as one pass/fail for the entire run, so a failure in step 4 doesn't hide what actually happened in steps 1 through 3.
Where Collections fit
Collections are for a workflow you'll run more than once — a login-then-fetch sequence you want to re-verify after a backend change, for instance. For a one-off "does this still fail" check, Replay is the faster tool.