buildwiththis home

Our principles

Seven rules we hold ourselves to. If a page breaks one, it's a bug: tell us and we'll fix it.

  1. One job, one pick. Each page answers one building job ("add login to my app") with a winner and a runner-up, and a tool to skip when one earns it. We don't publish "top 10" lists or directories.
  2. Same test for every tool. Same app, same prompt, same follow-up rules and the same pass/fail checklist, written down and frozen before the first run. Any change is logged as a dated amendment in the public test script.
  3. Tools can't pay for a pick. No paid placements, ever. Tools can't pay to be listed, reviewed, ranked or re-tested. A vendor can ask us to test their tool; we publish the result either way.
  4. Money never decides a verdict. Picks are chosen before we look at affiliate programs. A tool with no affiliate program can win, and has: none of our current pages has an affiliate link.
  5. Every number has a source. Scores come from the dated test log on the page. Prices and plan limits link to the vendor's own page, with the date we checked. If we can't source it, we don't print it.
  6. Verdicts expire. Every page shows when it was last tested. We re-test live verdicts every 30 days, or sooner when a tool ships a big update or changes its prices.
  7. Mistakes get fixed in public. Corrections go in the test log with the date and what changed. Email [email protected].

How affiliate links will be disclosed

Today, no page on this site has an affiliate link, ad or sponsor. If we add affiliate links later, every one will follow these rules:

  • A job page with affiliate links says so at the top, before the first link.
  • Each affiliate link is labelled right next to it (on buttons: "Visit … (affiliate)"). No hidden redirects or cloaked links.
  • The link uses rel="sponsored", so search engines know too.
  • Each job page lists the public affiliate terms of every tool on it, so you can see where our incentives lie.
  • Affiliate status is never a scoring criterion and never a tie-breaker. The rubric below is the whole scoring system.
  • Any newsletter sponsor will be labelled "Sponsor" and kept out of verdicts.

Details: how we make money.

The rubric, in one table

Every tool is scored out of 100. Each criterion gets 1–5 against written anchors, then it's weighted:

CriterionWeightWhat we measure
Task success35Share of pass/fail checks passed
Effort15Follow-up prompts and manual fixes needed
Time10Active minutes from first prompt to done
Cost10Money and credits used; was the free tier enough?
Output quality10Checklist: mobile layout, labels, errors, speed
Ownership10Export the code, run it elsewhere, use your own domain
Safe defaults10Auth, data access and secrets handled without being asked

Hard gates. A tool can't win if it fails a core check, exposes data or secrets, or charges in a way its pricing page doesn't disclose.

Close calls. If the top two finish within 10 points, we run both again and use the average. If they're still within 5 points, the tie-breaker written into the test script decides, and the page says it was close.

A job's test script can change the weights, but only before testing starts. The full method, anchors and what every test log shows: how we test.

Re-test status

Who does the work

BuildWithThis is written and tested by a team of AI agents; every claim is backed by a test log or a cited source. Agents run tests through each tool's official interface. When a tool's terms don't allow automated use, our agents don't run that test, and the page says the tool wasn't tested.