How TetherClaw Was Built

On Sunday April 4th, I walked into Best Buy with my pregnant wife and two kids — a one-year-old and a two-year-old. I was looking for a Mac Mini. A guy in the computer section told me someone had just walked out with the last one, and they couldn't get another until summer.

My wife was there. We talked about it. I bought a Mac Studio.

It was the first Mac I had ever owned.

The Terminal

I downloaded OpenClaw and set it up using my free Claude account to help along the way — which made things more complicated than they needed to be. It was my first time in the terminal, and I spent a lot more time there than I expected.

Between working as a certified welding inspector during the day and two toddlers in the evenings, I spent the next week setting up OpenClaw — wiring up a trading system, working on side projects. The setup took longer than it should have. I didn't care. I was hooked.

The Problem

The frustration built slowly.

During the day — on job sites, between inspections — I'd be on my phone. Making decisions with my AI agents, working through problems, pushing things forward. Then I'd get home, open my desktop, and a lot of that context wasn't there. I'd have to rehash decisions we'd already made. Re-explain where things stood.

Same thing in reverse. Build something at night at the desktop. Next morning on the phone — that context wasn't there either.

I was the sync. Manually carrying information between two interfaces that had no idea the other existed. The AI was useful. But it didn't remember me. Every new session, we started over.

April 16th

On Thursday April 16th, sitting at my desktop, I asked if there was a way to stop repeatedly going over everything we'd talked about during the day.

That question opened a rabbit hole. And somewhere in that conversation, the idea clicked.

This wasn't just my problem. Every person running AI agents had the same gap. The context breaks when you walk away from your machine. Nobody had built a clean solution for it.

I didn't want another messaging app. I wanted the exact same conversation, picked up exactly where I left off — on any device, at any time.

TetherClaw was born from that.

The Build

After that it was: work during the day, kids in the evenings, late nights after everyone was asleep.

I kept maxing out my Claude tokens. Upgraded to the $20 plan. Kept maxing that out. We were not stopping.

The app was submitted to Apple at 12:31 AM on April 20th. It wasn't ready to launch — not even close — but that was never the point of submitting it. The point was to keep building, and to build it in the right direction.

The difference now was I was using the app itself. Talking to my agents on my phone on the way to work. Picking up at the desktop at night. Stepping away for the kids and right back at it. The thing we built was solving the exact problem we built it to solve.

The Birthday Party

The moment I knew we had something was at my kids' birthday party.

Their birthdays are ten days apart, so we celebrated them together. People were starting to show up. I got a message — Hermes was ready. I snuck away for a few minutes to get an API key and finish the setup.

A little while later I got another message: reload TetherClaw and Hermes should be there.

I opened the app. Hermes Agent was sitting right there next to my OpenClaw agents.

That was the moment. Not a feature shipping. Not a metric. Just — yeah, this is something.

Why It Matters

Every person running AI agents has the same problem I had. The context breaks. The conversation doesn't follow you. You end up being the bridge — manually carrying what the technology should carry.

TetherClaw exists because that's a solvable problem. We solved it for ourselves first. Now it's available for anyone running agents who's tired of starting over.

The product didn't come from a business plan. It came from a real problem I lived every day.

Founder, TetherClaw