Disclosure first: this post is from the Checkcraft team (we make an API-checking desktop app), and the practice API below lives on our site. It is free, needs no sign-up and doesn't require our app. Postman, curl or the browser are enough.
Why another practice API?
JSONPlaceholder, Reqres and Restful-Booker are great for learning how to send requests. But they behave correctly, so they never teach the habit that matters most in API testing: a 200 status tells you the server understood you, not that the result is right.
So we made a small online-shop API where everything returns 200 and ten bugs are planted in the data.
Try one
GET https://checkcraft.ru/practice-api/v1/orders/5001.json
"subtotal": 3170.00,
"discount": 300.00,
"total": 2970.00
A typical test asserts status == 200 and total > 0. Green. But 3170 − 300 = 2870: the customer was overcharged by 100.
The other nine bugs are left for you. They are the kinds we keep meeting in real projects: a number sent as a string, a typo in a status, an extra field holding a secret, a list and a card that disagree, an answer that contradicts the log.
How to play
- Open the docs: https://checkcraft.ru/en/guides/api-testing-practice.html?utm_source=devto
- Send requests from the page, Postman or curl.
- When you spot a mismatch, pick the response, type the field and press Check: the page tells you if you found a planted bug.
The lesson for automated tests
If your test only checks the status and that fields exist, it would pass on all ten bugs. Check relationships instead: total = subtotal − discount, status agrees with paidAt, the order in the response matches the record in the log (/practice-api/logs/orders-service.txt). Whatever tool you use, feed your test a deliberately wrong response once. If it stays green, it checks nothing.
Found a bug we didn't plant? Tell us in the comments.
Top comments (0)