Ines Varela

Reviews

1. Lovable13 August 2026

   The prettiest first draft, and a metered second draft

   We ran all six briefs through Lovable three times each. It completed five of six on the first attempt, which nothing else in the index managed, and the booking flow it produced passed our accessibility audit with two minor focus-order fixes. The turn comes at brief four, the one with row-level security. Across three runs the agent needed eleven, fourteen and nine prompts to get the policy right, and each failed attempt cost credits. That is the whole review in one paragraph: outstanding at the part everyone demos, expensive at the part everyone actually ships.

2. v013 August 2026

   The only output I did not have to reformat

   I audit every generated interface in this index against WCAG 2.2 AA. v0 is the only tool that produced correct focus management on a modal without being asked, and the only one whose Tailwind output I would have merged as written. Across the six briefs it scored highest on markup quality and contrast. It is also the only tool where my notes are almost entirely about what it does not do rather than what it does badly: no database, no auth, no server logic worth the name.

3. Zite13 August 2026

   A very good site builder wearing an app builder badge

   On the marketing brief Zite was second only to Trickle on visual quality, and it beat every full-stack builder except Lovable and Totalum on our SEO and GEO axis: clean semantic markup, correct heading order, canonical control, fast pages, no client-side rendering of primary content. On the application briefs it simply cannot compete, and the absence of code export means there is no exit. My three stars is what happens when a tool is excellent at a third of the job this index measures.

4. Trickle13 August 2026

   Taste, on a platform that cannot carry an app

   On the marketing brief Trickle produced the only page all five panellists described unprompted as attractive. Type scale, spacing rhythm, restrained motion, and it does not look like the same template every time, which is more than I can say for most of this index. Then I tried the application briefs. There is no relational model, no roles, no export. My own dissent, published on this page, is that a 2.9 scalability score reads more generously than the reality deserves.

Comments

No comments yet.