About

I like building things. Usually tools that solve problems I keep running into.

What I’m Working On Now

DollhouseMCP

I built the core of DollhouseMCP in the summer of 2025, creating Personas, Skills, Templates, Agents, and Memory Elements as the core of a natural language toolkit for any LLM. It started because I had prompts scattered everywhere—Reddit posts I’d saved, Discord snippets, random notes. The simple solution would have been an organized folder. Instead, I built an MCP server. It works with Claude, ChatGPT, Gemini, and other MCP-compatible systems. Open source, AGPL licensed. You can see the Reddit announcement from August 2025.

Merview

Merview came out of frustration. I was sharing Markdown documents with a friend while we were working on DollhouseMCP, and we were using Mermaid diagrams to understand the workflow of the whole system. The existing tools were a pain—VS Code extensions that broke, online viewers that wanted accounts, SSO that didn’t work across devices. GitHub renders Mermaid fine, but only after you push. I just wanted to drop in a file and see it rendered.

I built it as a side project. First working version took a day or less. Then about a month—almost exactly—to get it stable and comfortable for anyone to use. It’s entirely client-side. Also open source, AGPL licensed, and still available now.

How I Think About Building

I treat tools like Lego blocks. The interesting work is figuring out how data flows between systems—where the connection points are, which APIs can talk to each other, what you can build when you snap them together in ways nobody’s tried before.

Hackathons were a great way of working like this. AT&T at CES, TVOT, MassChallenge—the wins were nice (grand prize + 4 category awards at one), but the real value was learning by getting in the weeds on things very fast.

At a hackathon, no one is your boss and you’re not anyone else’s. No one works for you and you don’t work for anyone else. You’re all working on a project together—it’s pure collaboration. If you think something is important, you need to explain why. If you can’t do it yourself, you need to convince someone—who has no reason to listen—why this matters more than what they want to do. Limited time, limited resources, no extra cash to just buy your way through problems. So you have to be clever and efficient with your ideas.

You get to know someone quickly when you’re working through a few mostly sleepless nights solving problems together. Some of the people I met at hackathons became partners on later projects, and friends I’ve stayed connected with for years. You learn fast who’s smart, helpful, and good to work with—and who isn’t. And, the prize money paid for the CES trip to Vegas more than a few times. :-)

Building with AI Tools

I get a lot of ideas for projects. I’m sure it’s some variation of ADHD. One of the hardest things to do previously was choosing what to focus on, and I got pretty good at it reigning in the part of my brain that wanted to work on the big thing AND have these other things going on.

I use AI tools for development now. Predominantly Claude Code and Codex CLI with DollhouseMCP and other MCP servers providing all kinds of extra capabilities as needed. I understand what they can do very well and I built tons of automation systems that are reliably programmatic, not based solely on AI semantics. Now I can take on more projects, and build side projects to make the tools to help me build more things—like a carpenter building jigs to make work easier on the next thing.

One of my first jobs was as a scenic carpenter at the Huntington Theatre in Boston. I knew carpentry, but effectively knew nothing compared to the senior carpenters there who’d been doing it for 30 years. One of them told me something I still remember: “If you need to do something once, you’ll probably need to do it again. Build a jig, make it adjustable, and label it so you can find it later.” Best Coding advice I ever got!

On Open Source

I’m a big believer in AGPL and copyleft licensing. MIT and similar licenses too often become a way for well-funded companies to extract value from developers without giving back to the ecosystem. I’d rather build things that stay open.

I’m not averse to making money. People need to be able to make money for the hard work they’ve done. AGPL allows commercial dual licensing—you can license the same code commercially while keeping the open source version protected. You can’t get that same benefit from MIT, not in the same way.

Both DollhouseMCP and Merview are AGPL, and more open source projects are on the way. Please feel free to work with them and share your improvements.

If you don’t want to share your modifications or can’t do that, we can talk about commercial licenses. There’s a deal to be made for anybody, whatever your process is—and either way, it helps support more value for the community.

Background

I’ve run web projects that millions of people have used; from New Girl, American Idol, and So You Think You Can Dance to live streaming for the Alliance of American Football, broadcast on TNT and CBS, with over a million live viewers on multiple web streams simultaneously. I was proud to lead the team that put it together in less than five weeks. Previously some of the same team members and I built Tomorrowish, “the first social media DVR,” with partnerships at truTV, FOX, E!, and Hulu. We were granted four patents in NLP and social media sync.

I’m not a trained software developer. I don’t have a degree in CS. Probably the thing I have the most experience with is product; understanding the pieces and parts of how something will be used, and more importantly, knowing who to ask to do something the right way.

AI tools now let me do a lot more building without having to bring in outside engineers for every piece and part. That doesn’t mean I don’t need them, and I do, often.

Part of why I’m building these things is to make it possible for people who are less skilled on the engineering side to craft their product, their ideas, their designs, their process—and let AI tools do the heavy lifting on implementation. But also to give engineers tools that can guide them on the product side, the design side, the copywriting side, the marketing side, if they choose.

And, the stuff I’m building now is giving me the most fun I’ve had in years.