There are too many things.

How It Works

1. Choose Free or Start
2. Enter your first name and email
3. Paste Prompt 1 into a fresh AI chat
4. Guided setup of your tools
5. Answer a few questions about you and your goals
6. Build your docbase
7. Solve one real problem
8. Keep going

Life's busy. It's complicated. Everything changes all the time. You need things, other people need things. There are always too many things.

This has been especially true for me lately. In the midst of all the usual life stuff, I've been working on a complicated dream of my own: building my first app. While I have a lot of experience with Excel, that's pretty much it for programming, so it's all newish to me. That's why I call myself a Field Coder.

I started with the free AI chat on Google. I designed something useful and amazing and totally outside my capabilities. When I tried to start actually building it, instead of just talking about it, everything fell apart. The AI forgot things, I forgot things, it was a mess.

I was really upset when that happened the first time and I had to start over. I was a little annoyed when it happened the sixth time. A pattern like that is just a problem to solve, and it made me wonder...

How can I use AI to do something complicated and actually useful over time?

Markdown-Driven Development, MDDD, is the method I'm actively developing to answer that question. So far it's working really well for me. I have some early but serious database architecture complete and tested, plus a plan for the next phase of work. And I actually know what's going on.

I also stopped in the middle to build this website, create the whole MDDD package you can access right now, and write these words (yes, I wrote them myself one keystroke at a time).

MDDD itself isn't complicated. It's centred around a docbase (a folder on your computer) that holds your goals, ideas, and the things you build, in plain text (or code, or whatever), that gets more useful as you use it. A side-effect is having an AI that actually knows what you're doing and can genuinely help you get it done.

There's a basic back and forth to the process. You discuss the thing you're working on with an AI chat which has no access to your docbase. You figure out the idea, shape, and contents of what you actually want to build.

I use claude.ai and pay for a subscription. That gives me access to Claude Code, which can interact with files in my docbase. If you're using a free AI, you manually do whatever Claude Code would do.

Maybe you want to build a form for your business, a website landing page, or a whole app. It doesn't really matter what it is, but the more complicated the project the more pieces there are. When you're done designing each piece, the AI provides you with clear instructions or writes a prompt for Claude Code.

Claude Code is different than regular AI chat, because it's meant for executing work. I've changed dozens of files simultaneously with different content for each without errors, and believe me I check.

I read every prompt, I look at every file, I do it every time. I catch mistakes, I fix mistakes, I move on. The scale and speed that the AI gives me is worth the upkeep. The compounding negative effects of those little errors, if left unfixed, makes the upkeep priceless.

I've tried free Gemini, free ChatGPT, and free Claude. Claude's the product I decided to pay for. But MDDD works with any of them.

When you get into the flow of MDDD, you might find that you start spinning ideas out all over the place. It happened slowly for me at first, but now I'm an idea factory. Some are good, more are bad. But all are worth considering, and a few are worth recording.

That's really the beating heart of MDDD and the docbase. It's memory of ideas you've had, decisions you've made, and the things those decisions created. It's also pretty cool to see it all happen over multiple sessions, with changing goals and priorities, and still know what's happening.

You can't build anything without a structure. All that great stuff you want to make needs something to stand on, something to hold it up, something to stop it from collapsing.

MDDD is that structure. Everything built within it is you.

My friend Michael has been using the first MDDD prototype since March 27, 2026. It took him about two hours to set up, and while I was there with him I offered minimal assistance. I designed the onboarding process to be self-unpacking, as I call it.

He ended that first session with an equipment checklist for his business that he could use the next day, a goal to make more, and an idea to turn them into a simple app for his employees. A few days later he started planning to document his estimating process, so he could teach someone else to do it.

The morning after that first session Michael sent me a text. He wanted to tell people about MDDD. That's when I switched gears and built this website, using MDDD, while developing MDDD, while building an SQL database.

I couldn't have done any of that on February 2, 2026, when I first started focused work on my app idea. I haven't collapsed in weeks, and I don't think I'm going to again.

But if I do, me and MDDD will find a way to fix it.

Free — $0

For the curious do-it-yourselfers.

What you need:

What you get:

Free — $0

Start — $1

For the productivity-focused.

What you need:

What you get:

Start — $1

Build — $5 (Coming Soon)

For Field Coders who want to do more with more.

What's planned:

Ask for Build

Build something today. Keep going next week.