How we test
A post on DevToolPicks makes a first-hand claim only when we can show where it came from. This page explains what that means in practice. It applies to posts from October 2026 on; older posts are being reviewed against it.
What a test is
We pick one task a reader actually has, such as deploying the same app, taking the same payment or running the same coding task. Then we do it on every tool in the comparison, under the same conditions. For each test we record:
- the date, the tool versions and the plan or tier
- the task and its inputs, so the test can be repeated
- what we measured, with units: time, cost on the bill, limits we hit, errors
- what happened, including failures and retries
- screenshots from the run
- what we could not test, and why
Accounts and money
We test on real accounts: free tiers and trials first, and paid plans when the comparison needs them, paid for by us. After a test we cancel trials and delete test data where the tool allows it. When a vendor gives us an account, credits or early access, the post says so.
What you see in a post
A post that compares or reviews tools starts with a How we tested box: the date, the versions and plans, the task, what we measured and what we left out.
Screenshots in the post come from our own runs. Numbers we measured match our test notes, and a number we worked out shows its arithmetic. Anything we did not test ourselves, such as a vendor's documentation or another site's benchmark, links to its source and is never described as our test.
Prices
Every price in a post links to the vendor's pricing page, with the date we checked it. We check the pricing pages of the tools we compare most often every week, and when a price changes we update the post with a dated note. We work from Europe, so some pricing pages show us euros. When a page would not show US dollars, the post says so.
When a test gets old
Tools change fast. A test older than 90 days does not back a new claim: we run the part that matters again before we update the post.
What we do not do
- We do not describe testing that did not happen.
- We do not compare tools we could not test. When one tool in a comparison could not be tested, the post marks it and gives the reason.
- We do not rank tools by affiliate commission.
Our editorial policy covers sources, corrections and sponsors. Spotted a test result that no longer matches what you see? Email [email protected].