Launching today
Agentic Website Builder 2.0 by Lokuma
Design, build, and run your site with a design agent harness
39 followers
Design, build, and run your site with a design agent harness
39 followers
Lokuma 2.0 is a design-aware agent harness for websites. Most AI builders can generate a first draft. But real sites need structure, taste, brand consistency, editing, publishing, forms, and ongoing updates. Lokuma connects planning, design, style, assets, site state, edits, and publishing into one agentic workflow β so your website feels designed, not just generated. Design, build, and run your site with agents.








Agentic Website Builder 2.0 by Lokuma
Hey Product Hunt!
Tech lead at Lokuma here.
The core shift in v2.0: we replaced the fixed "generate-once" pipeline with a real agent loop β the model plans, writes code, inspects output, and self-corrects until it's satisfied. That's what makes the quality jump feel so dramatic.
Β Β Β
A few things under the hood we're proud of:
1.Plan-first β the agent drafts a build plan before writing any code. Users approve it before execution starts.
Β Β Β
2.Targeted patching β edits use find-and-replace on the specific changed section, not full file rewrites. Faster and more predictable.
3.Live preview β updates as the agent works, not just at the end.
Β Β Β
We built this because we were tired of AI builders that look great in demos but break the moment you customize. Lokuma 2.0 is meant to be a real collaborative builder.
Happy to answer questions about the architecture β looking forward to your feedback!
The concept of a 'design agent harness' is intriguing. Does the agent iterate based on high-level feedback (e.g., 'make it more professional'), or does it require specific UI instructions?
Design Agent by Lokuma
@rivra_devΒ Actually both,but high-level feedback is the interesting case.
"Make it more professional" gets resolved against the design system the agent built first: type scale, spacing, palette, hierarchy. So vague intent becomes concrete moves inside an existing grammar, not a fresh guess.
Specific UI instructions work too, and the agent will flag when one breaks the system rather than silently complying.
The harness exists so the agent has enough context to interpret intent. That's the whole point.
Triforce Todos
Great one team, BTW, if someone eventually wants to hand this off to a developer or export it cleanly, what's the code quality like?
Design Agent by Lokuma
@abod_rehmanΒ Thanks! Yes, full export anytime.
It's agent-written code β modern stack, readable, componentized. Not hand-crafted by a senior engineer, but clean enough for a developer to pick up without a rewrite. The targeted-patching approach in v2.0 actually helps here: edits stay localized, so the codebase doesn't drift into spaghetti over many iterations.
what if I'm an agency building 5 different client sites?
Can the agent learn from one site's design system and apply lessons to the next one, or is each site starting fresh?
Design Agent by Lokuma
@boyuan_deng1Β Great use case, and one we hear from agencies a lot.
Right now each site is its own world β the agent has full memory within a site (brand, structure, iterations, live code), but nothing carries across. That said, you can still enforce consistency today through visual templates and instructions β set the design language once, reuse it as a starting point.
An agency-level layer where taste and components travel across client work natively is something we're actively designing. If you're open to chatting, we'd love to hear how your team would want it to work.
Design Agent by Lokuma
Hi Product Hunt,
I'm Mu, back again.
In March we launched Lokuma Design Agent here - a design intelligence layer that other AI agents call. It took #1 of the day, and that response shaped what came next.
Design Agent is still one of our core. It's evolving fast β the world is moving toward agent-first stacks, and design intelligence for AI is a long road we're committed to.
But there's another half of the picture.
A lot of people aren't building with their own AI agents, yet.
They're sitting down to make a website themselves.
They still need better tools.
So today I'm launching Lokuma Website Builder 2.0.
Design Agent is for AI.
Website Builder is for the people AI builds for.
Same conviction, different side of the screen.
Why now?
Over the past few months, agent architecture has matured. Tool dispatch, persistent memory, self-repair, observability β the runtime is finally ready. And SMBs and creators are entering AI website building in waves. But most of what they get is still a v1.
Most AI website builders generate a great first draft.
Then they leave.
You change the headline, the design breaks.
You add a section, the brand drifts.
The hero photo is wrong, but you can't say "make it more golden-hour" β because the agent doesn't remember what your brand looks like.
That's not a prompt problem.
That's a runtime problem.
So in the month after Design Agent shipped, we went heads-down to build the runtime underneath.
Lokuma 2.0 is a design-aware agent harness for websites.
The same agent that builds your site can edit it next month, restyle it next quarter, swap a palette, fix a broken form, ship the change. It remembers your brand, your structure, your previous iterations, your live source code.
Generation isn't running a website.
A first draft isn't enough.
You need something that stays.
Why us?
I've spent the last decade building tools designers actually use β Readdy and Creatie reach 500,000+ designers and creators today.
Lokuma Website Builder 1.0 quietly shipped in February and showed us where AI-built websites actually break.
Design Agent gave AI a designer's brain.
This one gives that brain a body β and that took the whole team.
We're a small indie crew β designers who write code, AI researchers who care about typography, growth gurus who were running AI-native marketing experiments before that was a category. The roles blur, deliberately. A harness like this lives in the seams between design and runtime, between agent loops and SMB workflows β exactly where mixed-discipline teams move fastest.
Curious how others here see it:
When your AI ships v1, what's the first thing that breaks the second time you try to change it?
β Mu
Design Agent by Lokuma
I've spent the last few years marketing AI website builders. From inside that seat, the pattern is hard to miss: every new launch promises something different, but the outputs converge within a quarter. Different brands, same skeletons. Different prompts, similar typography. After a while it stops being a tooling problem and starts feeling like a category problem.
The thing that's hard for AI isn't generating a page. It's generating constraint. Design is mostly about what gets left out β the hierarchy that decides what earns attention, the system that decides what's allowed at all. Models are trained on abundance; they want to add. Design wants to subtract. That's the real gap.
Lokuma 2.0 starts there. It builds the design system first β color, type, spacing, hierarchy, the whole grammar β before a single page exists. The site is then composed inside the system, not the other way around. Drop in a reference site and it picks up the design language without copying the layout. Bring an old site across and it gets rebuilt on top of a proper system, content and assets intact.
It's a strange thing to ship from inside a category I've watched commoditize itself. But this is the part that's hardest to fake, and the part that β if it works β holds up. AI is going to keep making generic things faster. Taste is the only layer that compounds.
Agentic Website Builder 2.0 by Lokuma
Hey Product Hunt π
Joy here, back for round two.
In March, we launched Lokuma Design Agent β a design intelligence layer built for other AI agents to call. It hit #1 Product of the Day, and what we heard from that community changed everything about what we did next.
Design Agent is still core to what we're building. It's moving fast, agent-first stacks are becoming the default, and we're in it for the long game on design intelligence for AI.
But there's a side of this we hadn't fully addressed.
As a branding & marketing agency owner, I've seen how SME owners struggle.
You're wearing ten hats. You're good at your craft β whether that's running a restaurant, a studio, a clinic, a shop. But building a website? That's a different skill set entirely, and hiring someone to do it right isn't always an option.
I watched businesses around me struggle with this for years. Not because they lacked ambition, but because the tools assumed too much, too much time, too much budget, too much technical fluency. They'd settle for something that looked cheap. Or they'd spend money they didn't have on something they couldn't maintain.
That gap is what Lokuma is really about. Not just a better AI tool, but an honest attempt to give small businesses something they've never quite had: a website that looks like it was made for them, and actually stays that way.
What Lokuma 2.0 actually is:
Most AI builders are one-and-done. They generate, then vanish. What you're left with is a site you can't confidently touch β because the thing that built it isn't there anymore.
Lokuma 2.0 is different. It's a design-aware agent harness: the same agent that builds your site sticks around. It comes back next month when you need to update the headline. Next quarter when you want a full restyle. When a form breaks, when the palette needs a refresh, when your offering changes and the whole page needs to catch up.
It knows your brand β not as a setting you configured once, but as something it's held onto: your visual language, your structure, your previous iterations, your live source code. Every edit is informed by everything that came before.
Because here's the thing most builders miss: generating a website isn't the same as running one. A great first draft is just the beginning. The real test is what happens the second time you need to change something β and the third, and the tenth.
You don't just need something that launches. You need something that lasts.
So join us and explore Lokuma 2.0 together.