Inspiration

I started building Unplain because I was tired of how unnecessarily complicated creating a decent-looking PDF could be. As a designer and marketer, I've worked on plenty of ebooks, lead magnets, and reports. The process was almost always the same: write the content in Google Docs or generate it with AI, move everything into Figma or another design tool, spend hours formatting it, and then repeat half the work whenever the content changed. I once had to rebuild a 20-page document just because we changed the font. Another project took months to get a relatively simple PDF published. With AI making content creation faster than ever, this workflow started to feel particularly ridiculous. We can generate an entire ebook in minutes, but turning it into something presentable still takes hours. That's where the idea for Unplain came from.

What it does

Unplain turns existing content into polished, branded, publish-ready PDFs in minutes. You can upload a Word document, paste Markdown from ChatGPT or Claude, or connect your AI assistant directly through MCP. Unplain takes care of the design: typography, layouts, page breaks, covers, tables of contents, branding, and more. You can customize everything in a visual editor or simply export the result. The goal isn't to replace writing tools or professional designers. It's to eliminate the repetitive design work that happens between having finished content and having something ready to publish.

How I built it

I'm a solo founder with a background in graphic design, marketing, and development, so Unplain has been a combination of all three. I built the application using Next.js, React, TypeScript, and a PostgreSQL database, with Stripe handling subscriptions. One of the more interesting parts was building the document layout engine. Unlike a typical design editor, Unplain needs to take unpredictable content and automatically distribute it across pages while maintaining consistent spacing, typography, and visual hierarchy. I also built an MCP integration so AI assistants can interact with Unplain directly. Instead of copying content between applications, you can ask your assistant to create a PDF and have it generated through Unplain. AI coding tools have been a big part of my development process, allowing me to move quickly as a solo developer while spending more time on product design and the overall experience.

Challenges I ran into

The biggest challenge has been making automatic document design actually work. It's relatively easy to put text on a page. It's much harder to make sure that a 30-page document looks consistent, headings don't end up stranded at the bottom of pages, tables don't break awkwardly, and everything still works when users change fonts, margins, or content. Another challenge has been finding the right balance between automation and customization. I wanted Unplain to be simple enough that someone could create a PDF in a few clicks, but flexible enough for marketers and designers who care about their brand identity. And, of course, building a product as a solo founder means constantly switching between development, design, marketing, customer support, and everything else.

What I learned

Probably the biggest lesson has been that automation is only useful when the result is actually good enough to use. Saving someone 30 minutes doesn't mean much if they still need to spend another hour fixing the output. I've also learned how important it is to get people to actually use the product rather than just building more features. Some of my assumptions about how people would interact with Unplain turned out to be wrong, and watching real users has been incredibly valuable. Finally, building Unplain has made me much more interested in the relationship between AI and design. I don't think AI needs to replace creative decisions. There's a lot of value in using it to handle repetitive work while leaving people in control of the things that matter.

What's next for Unplain

I'm currently working on making Unplain more automatic, with smarter design decisions, better templates, and less configuration required from users. I also want to expand the MCP integration and make it easier to create documents directly from AI conversations, including on mobile. The long-term vision is pretty simple: going from finished content to a professionally designed document should take minutes, not hours. And ideally, you shouldn't have to think about formatting at all.

Built With

Share this project:

Updates