DEV Community: Michael Wang

How 6 Theme Families Replaced 1,000 Random Outputs

Michael Wang — Sun, 22 Mar 2026 17:57:00 +0000

Every AI website generator has the same problem: the output looks random. You prompt for a bakery site and get something that could be a SaaS landing page with a different color. The architecture underneath doesn't actually understand type.

I spent two weeks rebuilding how CapsuleWeb resolves visual identity, and the core decision was replacing open-ended style generation with six curated theme families.

The Six Families

clean-editorial — whitespace-forward, type-heavy, minimal color
modern-product — sharp contrast, geometric, tech-native
soft-playful — rounded corners, warm palette, approachable
boutique-warm — textured, earthy, small-business feel
utility-local — dense, information-first, no-nonsense
elegant-dark — dark backgrounds, refined type, high contrast

Each family defines font pairings (Outfit, Fraunces, and others properly wired through the renderer), color logic, spacing ratios, and component behavior. When a user types a prompt, the system runs resolveThemeFamily(archetype + vibe + color) to match intent to family before any layout generation starts.

Why This Matters Technically

Before this, the 586-line monolithic renderer tried to handle every visual decision inline. I broke it into 11 focused components, each aware of which theme family it's operating under. A button in boutique-warm has different padding, border-radius, and hover behavior than the same button in utility-local.

This isn't cosmetic. It's structural. The theme family constrains the decision space so the AI generates coherent pages instead of statistically average ones.

The Gradient Problem

One specific fix: I killed gradient sludge across all outputs. Earlier generations loved throwing muddy linear gradients on hero sections regardless of context. Now gradient use is family-gated — elegant-dark gets subtle gradients, clean-editorial gets none, soft-playful gets them only on accent elements.

What Changed in Output Quality

A restaurant prompt now reliably lands in boutique-warm or utility-local depending on the vibe words used. A portfolio prompt resolves to clean-editorial or modern-product. The outputs feel like they belong to a category instead of floating in generic AI space.

The whole generation still takes about 30 seconds and costs $1 with no account required. But the visual coherence gap between this and what shipped a month ago is significant.

Constraining the design system made the AI better at its job. Fewer choices, better pages.

Deployment Latency Is a Solved Problem — Here's the Proof

Michael Wang — Thu, 19 Mar 2026 17:53:45 +0000

The Experiment

I wanted to answer a simple question: what is the actual minimum time between a user deciding they need a website and that website being live on the internet?

Not the theoretical minimum. The real, measured, production minimum.

The answer: 27 seconds.

Where the Time Actually Goes

I audited the deployment flow of every major website builder. Here's where time accumulates:

Step	Average Time
Account creation	47s
Email verification	30s–5min
Template browsing	2–8min
Template customization	5–30min
Domain/subdomain config	1–3min
Publish + DNS propagation	10s–48hrs

The actual compute — generating markup, writing files, configuring a subdomain, deploying to a CDN — takes under 10 seconds on modern infrastructure. Everything else is ceremony.

Removing the Ceremony

CapsuleWeb was my testbed for this thesis. The entire user flow is:

Write one sentence describing what you need
Pay $1
Receive a live URL on a capsuleweb.site subdomain

No account. No template picker. No editor. No DNS configuration.

The architecture decisions that make this possible:

No accounts — eliminates auth, verification, session management, and the entire "returning user" state machine
Subdomain-only deployment — no DNS propagation delay, no domain verification, instant routing
Single prompt input — no multi-step form, no configuration choices, no decision fatigue
Pre-warmed generation pipeline — the LLM context and deployment infrastructure are always ready

The Tradeoff

The obvious cost: no editing. Once your site is generated and deployed, that's it. There's no dashboard to return to, no account to log into.

This is a deliberate constraint, not a missing feature. The target use case is people who need a page live right now — an event page, a quick business listing, a resume link, a product landing page. For those cases, 27 seconds to live beats 27 minutes to perfect every time.

What This Proved

Deployment latency was never a compute problem. It was always a product design problem. Every step in a traditional builder exists because someone decided to put it there, not because the user required it.

Strip the steps that serve the platform instead of the user, and you discover that the internet already had the infrastructure for instant deployment. We just buried it under forms.

The site is live at capsuleweb.site. One prompt, one dollar, one live URL. I'm curious what the dev community thinks about the no-account tradeoff — is it a dealbreaker or a feature?