Skip to content

Hurl

Hurl runs API requests written in plain text and checks replies against rules you define. It gives repeatable pass or fail results for checks a teammate can review.

Visit Hurl

How does a request become a repeatable check?

Write the request in a text file, then write what must be true of the response. For a fictional shop, a request to read opening hours might pass only if the status indicates success and the hours field contains a time. Hurl can report a failed rule even when the service replied successfully.

Where can private values still appear?

Keep credentials out of the saved test file. Hurl documents redaction for values designated as secrets in logs and reports, but a raw response can still contain a secret. The JSON report can also save unchanged response bodies to disk. Review where a build stores test output and who can download it.

What travels with the test when someone else runs it?

Keep the plain-text test file and a list of the fictional inputs it expects. Hurl is an open-source command-line runner. The person running it supplies the executable, the correct test address and any separately managed credentials. A project or build service still pays for its own compute and test infrastructure.

What would make a successful-looking reply fail?

On paper, write a two-part check for the opening hours of a fictional shop. Both a success response and a valid opening time should be required.

  1. Write a passing reply
    Use a success status and the sample time 10:00.
  2. Change one detail
    Keep the same status and remove the time. This second reply should fail the content check.
  3. Name the failure clearly
    Describe the failure as “opening time missing.” Keep that message with the test so a teammate knows what to inspect.