A markdown viewer
for vibe coders.

PrettyMD is a simple, beautiful app for reading the .md files your AI writes — specs, plans, skills, agents, and more.

7-day free trial No sign-up
PrettyMD showing a rendered markdown document with inline comments and an outline
welcome.md — PrettyMD

Works with every agent

Claude Code logoClaude Code Codex logoCodex opencode Cursor logoCursor

Why PrettyMD?

Sure, you could open them in VS Code or on GitHub.

Here's what you'd be missing:

All your markdown, in one place.

Open a file or a folder, and see only the .md files inside — all beautifully rendered.

Leave comments your agent reads.

Highlight any line and leave a comment. Your agent picks it up, replies back to you right there, and makes the change.

See what your agent changed.

Your agent rewrites these docs all the time. PrettyMD highlights what's new since you last looked, so you read the lines that changed and skip the rest.

And everything else you'd expect.

Search, outline, themes, focus mode — every detail built for reading.

Sidebar
Your folder's .md files, listed down the side
Outline
Jump between headings in a long doc
Tabs
Keep several docs open at once
Light & dark
Warm light and dark themes
Focus mode
Hide everything but the words
Keyboard-first
Shortcuts for everything

Pay once. It's yours.

$19one-time

7 days free · no credit card · no sign-up

Everything included
  • Read every .md in your folder, fully rendered
  • Mermaid diagrams, code, and the full Markdown spec
  • See what your agent changed, and comment on it
  • Outline, tabs, and instant search
  • Light and dark themes
  • Mac and Windows
  • Free updates, forever
  • Fully offline

Why I made this

I never really read what my AI wrote.

When my agent handed me some big plan, I barely read it, honestly. The reading experience was just bad. Raw markdown in Notepad, or VS Code. So I'd skim for two seconds and tap "Implement the plan."

It got implemented, sure. But something was always off. It built something I never approved, or did the right thing in the wrong way.

For a while I blamed everything else. Then one day I sat down and actually read one of those plans, properly, for once. And it had SO many things in it. Stuff I didn't want, stuff that was ambiguous, calls it made on its own.

That's when it hit me. I had to take every markdown my AI writes seriously. Plans, skills, AGENTS.md, all of it. Actually read the thing, not just nod and approve.

But I wanted more than a nice renderer. I wanted to comment on what my agent wrote, and see a clean diff of what it actually changed instead of digging for it myself.

So I built the viewer I wished I had. That's PrettyMD. If it gets you reading what your AI writes, it did its job.

SomyaS
Cheers,
Somya

P.S. It's just me building this. So if PrettyMD makes your work better, or if something about it drives you nuts, shoot me a DM on X anytime. Praise, bugs, or pure venting, I read all of them.

Frequently asked questions

You get the full app free for 7 days. Just enter your email to start — no credit card required. Every feature is unlocked during the trial; nothing is held back.

The app locks until you buy a license. Your work is safe either way — your markdown and comments are plain files on your computer, so you keep them no matter what. A one-time $19 license unlocks everything again.

One-time. PrettyMD is a single $19 payment, not a subscription. Pay once and it's yours — no monthly fees, no renewals.

Because we haven't signed the app with a paid developer certificate yet. Windows (SmartScreen) and macOS ("unidentified developer") show that warning for any app without one — it's about the missing certificate, not anything unsafe. To open it: on Windows, click More info → Run anyway; on Mac, right-click the app and choose Open.

In a separate file right next to your markdown, named yourfile.md.comments.json. Your original markdown is never touched — comments live entirely in that sidecar file. It's plain, readable JSON, and you can hide these files from git with one click.

Two. One license activates on up to two of your computers — say, a laptop and a desktop.

Both. Because comments are stored in that plain .comments.json file in your project, your agent can read the ones you leave and write its own. Comments from an agent show up tagged with its name (like "Claude"), so you can always tell them apart from yours.