mdman

01 · The brief

Docs that don't rot.

mdman is a markdown editor for technical teams. Documents stay plain markdown, AI chat answers questions grounded in the document you're reading, Mermaid diagrams render inline, every edit is versioned and a workspace can be published as a blog on your own domain.

Its audience of engineers, technical writers and product teams has watched every documentation tool promise to fix the wiki and then become another wiki. They're allergic to marketing and they read the fine print: lock-in, privacy and where their keys go.

02 · What the page had to do

Earn the trust of people who hate landing pages.

The page had to get a developer to try mdman in a single visit. That meant it had to:

• Say what mdman is in one sentence, and why it's different from the wikis teams already use.
• Show breadth without bloat: many capabilities, each clearly useful.
• Answer the trust questions up front: portability, self-hosting, encryption and bringing your own AI key.
• Help different teams recognise their own use case.
• Keep the path to a first document free and friction-free.

03 · How we delivered

Written in the reader's own register.

The copy is short, dry and specific, the way developers write to each other: “Docs that don't rot.” “Nine features. No bloat. Each one earns its keep.” The claims are concrete, like Mermaid rendered on save, autosave on every keystroke and server-pushed edit alerts with no polling, because that's what this audience checks.

The design follows the same logic: dark by default, one orange accent, a single sans-serif and generous spacing. Nine features sit in a scannable grid, then deeper sections cover editing (“Markdown in. Rendered out.”), diagrams as source code, AI chat with your own Gemini key, and publishing a workspace to a custom domain with themes, TLS and SEO metadata handled.

Trust is addressed where the question comes up: plain-markdown export, a self-hostable stack, encryption in transit and at rest, and AI that runs on your key. A “one editor, six jobs” matrix, from API docs and engineering blogs to RFCs, internal wikis, meeting notes and roadmaps, lets each team find its reason to start, and seven FAQs cover the comparison and pricing questions.

04 · Built to be found

Ready for search and AI answers.

SoftwareApplication and FAQPage structured data describe the product and its answers to search engines, an llms.txt file summarises it for AI assistants, and a blog lays the groundwork for a content programme. The Next.js build is deployed on Vercel.

05 · What shipped

The deliverables.

• Positioning and landing-page copy
• Dark, keyboard-first visual design
• Feature, use-case and FAQ architecture
• Blog
• Structured data, sitemap and llms.txt
• Next.js build, deployed on Vercel