WeWeb has launched its AI features. The AI works well in some areas, but not so well in others. It’s interesting, but it doesn’t completely change everything.
This is my experience after using it on client projects.
WeWeb added AI features to its low-code platform. You can now prompt it to generate UI components, create workflows, set up backends (either WeWeb’s native Tables or Supabase), and build custom components.
They launched this as part of WeWeb 3.0 beta in early 2025, and it’s available right in the editor. There are no separate tools and no switching between platforms. Just type what you want and see what happens.
So, the UI generation with WeWeb AI works. Sort of.
You can screenshot a design from Figma or describe what you want, and WeWeb AI will create something that resembles what you asked for. But the issue is that it won’t match your design system.
I tried feeding it a client’s designs for a dashboard with their specific spacing, colours, and typography.
What I got back was functional. But the spacing was off, the fonts weren’t quite right, and the elements were not assigned any classes or typographies from the app’s design system.
One thing that improved: as of May 2026, the AI can now generate and configure 20+ native elements in context, including Datagrid, PDF Viewer, ChartJS, Timeline, Iframe, OTP Input, and more. Previously you’d get a placeholder that needed rebuilding manually. That’s no longer true for most common elements.
The design system problem is still unsolved though. WeWeb has said it’s a priority for mid-2026. Until it lands, the AI produces output that works but doesn’t look like the rest of your app.
Beginners may see it as magic. Those with experience know it’s a decent starting point that still needs cleanup. Bubble has its own AI offering too. Read our Bubble AI review for a direct comparison of how the two approaches differ.
Related: WeWeb vs Bubble
WeWeb added multi-page AI generation in May 2026. You can now describe an entire app and the AI builds across multiple pages, not just one screen at a time.
This matters because building screen-by-screen was one of the bigger friction points. Every new page was a fresh prompt with no context from what you had already built. The AI had no idea what your nav looked like or how your other pages were connected.
With multi-page, the AI maintains a shared nav, keeps layouts consistent across pages, and wires up interactions between screens. You still need to review everything, but the starting point is closer to a coherent app rather than a pile of disconnected pages.
The design system limitation applies here too. Consistent layouts, yes. Matching your brand tokens, not yet.
WeWeb AI can generate workflows. It knows when to use if-then statements, understands pass-through conditions, and even adds comments to explain its actions. That’s good.
The challenge is that it prefers to use Javascript in some places, creates dozens of single-use variables, and fails on complex logic.
And once it creates a workflow, you still need to double-check what it did or make minor modifications.
The AI was better when it first came out, but this new version seems to have more bugs.
That said, for simple CRUD operations or basic form submissions, the AI does fine.
This is where WeWeb AI impressed me.
Need a custom date picker with some specific checks? Describe it, and you’ll get something that mostly works. Want a complex data table with sorting and filtering? Same deal.
I have seen people use WeWeb AI together with Cursor to create components. This combination works well. WeWeb AI manages the basic structure, and Cursor improves the code. In the end, you get something that is usable.
It still has rough edges, though. Sometimes, the components don’t handle responsive design well or have conflicts. But compared to the UI generation, this feels reliable.
Since April 2026, WeWeb comes with its own native backend: WeWeb Tables. It includes a Postgres database, auth, workflows, and file storage. You no longer need an external backend to build a working app.
The AI can generate your WeWeb Tables schema from a description, create the tables, and wire them up to your frontend. For simple apps, this works well.
WeWeb AI can also set up a Supabase backend: schema, tables, RLS policies, and edge functions. This is the more polished option for anything complex. Supabase has been optimising for AI tools (Bolt, Lovable, etc.) and it shows.
I would not recommend either option for non-technical people without some guidance. WeWeb Tables is easier to start with. Supabase gives you more control but requires understanding data security and API design.
Related: WeWeb and Supabase
Here’s what would make this useful for agencies like mine:
WeWeb updated its pricing to factor in the tokens for AI. There are no AI feature restrictions on any plan, and the different plans have varying token limits.
Based on my experience, those AI tokens finish quickly.
I’ve seen people complain about burning through 500k tokens on simple requests. That’s not great.
The workspace plans start at $20/month (Essential, billed annually). Site plans start at $13/month for the Front-end only plan. You’ll likely need to upgrade to a higher workspace plan if you use AI frequently.
If you live in a few specific countries, you’ll see prices in your local currency and at a lower amount due to pricing parity.
The feedback has been mixed.
Some users love the UI generation for quick prototypes. Others are frustrated with token consumption and inconsistent results.
Use it, but don’t depend on it.
AI tools are useful. But they’re not magic.
Need a custom component? WeWeb AI can generate it. Want to prototype a workflow? WeWeb AI can give you the basic structure.
For production apps that require reliability and maintainability? I’m still doing most of the development myself.
It’s difficult to explain complex business requirements and dependencies.
WeWeb AI is a decent tool that’s getting better.
If you’re just getting started with WeWeb, it’s useful. If you’ve been using WeWeb for a while, it’s a useful intern.
Don’t expect it to build your entire app perfectly.
If you want someone who’s built 40+ WeWeb apps to take a look at your project, our WeWeb agency can help.
WeWeb AI gets you started. Knowing WeWeb gets it done.
We've built 40+ WeWeb apps. If you're evaluating WeWeb for a real product (not a prototype), a 30-minute call will tell you whether it's the right fit and what to watch out for.
Related: WeWeb and Xano
Himanshu runs NocodeAssistant, a development agency that builds internal tools and SaaS products for growing companies. He's worked directly with every client since 2019. Same person from kickoff to post-launch.
Connect on LinkedInWork with us
30 minutes with a team that builds WeWeb apps for real production use.
Continue reading
SAP and NetSuite are built for companies ten times your size. Here's why a $12,000 custom tool fixes more than a $150,000 ERP, and how to know which one you actually need.
Discover the cost of hiring a Bubble agency for your business. Find out what influences the price and see a real SaaS example by a Bubble agency.
Your team is doing $15/hour work while you pay them $80/hour. Here's what manual processes cost a growing SMB and which one to fix first.
