We Published 15 Landing Pages in One Hour. Here's What Happened.
The HTML Pub Team•We build landing pages for a living. We know what goes into them - the brief, the copy iterations, the layout debates, the "can we try one more version" conversations. A decent page takes 30-40 minutes on the fast end. A good one takes longer.
Last Tuesday, we sat down with Claude, our MCP connector, and HTML Pub. We gave ourselves one hour and a single goal: publish 15 landing pages for 15 different use cases.
We kept a timer running the whole time.
The Rules
We wanted this to be a real test, not a demo reel. So we set constraints:
Each page had to serve a distinct purpose with a different audience and goal. We weren't going to copy-paste the same template fifteen times and call it done.
We'd use two workflows: the first five pages through copy-paste (generate HTML in Claude, paste into HTML Pub manually), and the remaining ten through the MCP connector (publish directly from the Claude conversation without ever leaving the chat).
We'd time each one. And we'd be honest about which pages were actually good and which ones needed more work.
Pages 1–5: The Copy-Paste Workflow (Manual)
Page 1 - SaaS Free Trial Landing Page Prompt: "Build me a landing page for a SaaS tool offering a 14-day free trial. Target audience: marketing managers at mid-market companies. Emphasize time savings and integration with existing tools."
Time to live: 4 minutes 12 seconds.
The HTML came back clean. Copy, paste into HTML Pub, scan the preview, publish. The headline was stronger than what we'd have written on first pass - "Stop Managing Your Tools. Start Managing Your Results." We kept it.
Page 2 - Webinar Registration Page Time to live: 3 minutes 48 seconds.
Faster - we'd found our rhythm with prompting. We gave Claude the webinar topic, the speaker bio, and three bullet points on what attendees would learn. The page had a countdown timer, speaker headshot placeholder, and a clean registration form.
Page 3 - E-book Download Page Time to live: 3 minutes 15 seconds.
Page 4 - Product Launch Announcement Time to live: 4 minutes 30 seconds.
This one took longer because we went back and forth on the hero section. The first version was too feature-heavy. We told Claude to rewrite the top half with the "so what" framing - why should someone care about this launch? Second version landed.
Page 5 - Event Recap Page Time to live: 2 minutes 55 seconds.
The fastest one. Event recaps have a predictable structure and Claude nailed it first try.
Copy-paste workflow total: 5 pages in 18 minutes 40 seconds. Average: 3 minutes 44 seconds per page.
Pages 6-15: The MCP Connector Workflow
The MCP connector changes the workflow completely. You stay inside your conversation with Claude. You describe what you need and the page gets built and published without switching tabs or pasting anything.
Page 6 - Consultant Services Page Time to live: 2 minutes 10 seconds.
Almost a full minute faster than the manual workflow. The difference was all in the friction that disappeared - no copying, no pasting, no switching windows. Just describe what you want, say "publish," and it's live.
Page 7 - Nonprofit Donation Page Time to live: 2 minutes 35 seconds.
Page 8 - Course Pre-Launch Waitlist Time to live: 1 minute 52 seconds.
This was the one that surprised us. Under two minutes for a waitlist page with a headline, value prop, social proof section, and email capture. We stared at the live URL for a second before moving on.
Page 9 - Portfolio Page for a Freelance Designer Time to live: 2 minutes 40 seconds.
Slightly longer - portfolio pages need more visual structure, and we spent extra time on the prompt describing the layout we wanted.
Page 10 - Restaurant Catering Menu Page Time to live: 2 minutes 15 seconds.
**
Page 11 - Real Estate Open House Page** Time to live: 1 minute 48 seconds.
Page 12 - Fitness Challenge Signup Time to live: 2 minutes 05 seconds.
**
Page 13 - Local Business "About Us" Page** Time to live: 2 minutes 22 seconds.
**
Page 14 - Job Posting / Careers Page** Time to live: 2 minutes 30 seconds.
**
Page 15 - Coming Soon / Teaser Page** Time to live: 1 minute 38 seconds.
The last page was the fastest of the entire experiment. A teaser page with a countdown, one line of copy, and an email capture. Published in under two minutes.
MCP connector workflow total: 10 pages in 21 minutes 55 seconds. Average: 2 minutes 12 seconds per page.
The Final Tally
Workflow
Pages
Total Time
Avg Per Page
Copy-paste (manual)
5
18 min 40 sec
3 min 44 sec
MCP connector
10
21 min 55 sec
2 min 12 sec
Combined
15
40 min 35 sec
2 min 42 sec
Fifteen pages in forty minutes. Nineteen minutes under our one-hour budget.
What Surprised Us
2 minutes 42 seconds per page is fast. Obviously. But after the first few pages, the speed stopped being the interesting part.
What we kept coming back to was how it changed the decision about whether to build something at all. Normally, when we have a landing page idea, there's a cost-benefit calculation: is this idea worth 40 minutes? Worth pulling in a designer? Worth the revision cycles? Most ideas don't clear that bar. They go into a Notion doc and stay there.
When a page takes two minutes, that calculation goes away. You just make it. We found ourselves building pages we'd never have justified otherwise - the restaurant catering page, the freelance portfolio, the fitness challenge. The cost of trying was close to zero, so we tried everything.
Not every page was perfect. Pages 4 and 9 needed more refinement than the others. Page 4's first hero section was too feature-heavy. Page 9 needed a more specific prompt to get the layout right. But needing refinement at the two-minute mark feels completely different than needing it at the forty-minute mark. You still have energy to actually do the editing.
The MCP workflow averaged about 90 seconds faster per page than copy-paste. Over fifteen pages, that's roughly 15 minutes. The bigger difference was staying in one place the whole time - every tab switch during the copy-paste workflow broke our train of thought a little, and those interruptions added up.
What This Means If You're a Marketer
When every landing page costs 30-40 minutes and a round of stakeholder reviews, you only build pages for campaigns that have already been approved, budgeted, and scheduled. The quick experiment, the audience variant, the "what if we tried this angle" - those ideas die in your project management tool because the execution cost doesn't justify the uncertainty.
At two minutes per page, you can test ideas before they're fully baked. You can build the variant. You can create the niche page for the audience segment that's too small to justify a traditional build. If it doesn't work, you spent two minutes.
We finished fifteen pages in forty minutes and had nineteen minutes left over with nothing to do but look at what we'd made.
Try It Yourself
HTML Pub is currently in early access - visit www.htmlpub.com to try it out today!
The fifteen pages we built are still live. You can see every one of them. Nothing was polished after the experiment - what you see is what came out of the workflow.
HTML Pub is built by the team at Leadpages. We make tools that let you publish AI-generated landing pages in seconds.