Lately I’ve seen a wave of “AI for front-end” tools claiming they can handle everything — layout generation, component design, Tailwind formatting, even full landing pages with forms and logins.

I’ve been experimenting over the last 2 months with a few of them (Cursor, v0, and Skywork) and wanted to share some honest thoughts + hear from others who’ve gone down this rabbit hole.

Cursor: amazing when you already know what you’re doing. It’s like Copilot on steroids. It nails the boring stuff, splitting files, fixing state errors, formatting Tailwind, etc. Tab completion is way ahead of VS Code. But once you try to let it run wild with “vibe coding” or build something complex end-to-end, it collapses fast. I usually end up rewriting after a few prompts.

v0: very cool for quick component scaffolding. It’s fast and predictable, but the outputs can look a bit sterile. You still need a designer’s eye.

Skywork: this one, it’s not a coding assistant like Cursor or Claude, but a visual AI builder that generates a full web app (front + back). You can edit everything visually afterwards. I tried it for a multi-step signup + dashboard flow, and it actually got the UX hierarchy right and consistent styling throughout. Feels more like a “lovable AI dev” than a coder. Here’s my task link, which includes the webpage I generated. If anyone wants to learn how to prompt and generate a webpage using AI, feel free to check it out: https://skywork.live/share/project/1978361422294982656?sid=-slc9kPqal&t=

So far, I’ve found the combo of: → Cursor for code-level help → Skywork for full-page scaffolding / visual generation → and manual cleanup after ...to be way faster than starting cold.

Curious, what’s everyone else’s workflow? Anyone using AI for production-ready front-end work yet, or is it still “rewrite everything after”?