Blog
Building in public.
Practical notes on AI automation, agent frameworks, and building intelligent systems — from someone who ships production code, not slide decks.
Stranger Danger for My Own Agents: Wiring im-in-danger into Forge's Web Tool
I published a prompt-injection filter built on TypeSafe's Jev model, then went looking for where it belonged in my own agent runtime. The honest answer: one part ships now, the rest waits for a deployment that needs it. Here is the one seam it fits, what the runtime would have to build, and the bug I found on the way.
Read moreAI for Small Business: A Practical Guide
A practical guide to AI for small business — what it actually delivers, how to tell a builder from a slide deck, real use cases, and the questions to ask before you hire anyone.
Read moreWhy Small Businesses Should Build AI Agents (Not Buy SaaS)
Learn how to build AI agents tailored to your business instead of renting generic SaaS. A cost comparison and the 3-step process we use to build custom AI agents for small businesses.
Read moreStop Burning Tokens: What We Learned Replacing an AI Agent with a Dispatch Gateway
We replaced a multi-agent orchestration system with 400 lines of Python. Response time dropped from 30 seconds to 3. Cost dropped 100x. Here's what we learned about when agents make sense — and when they don't.
Read moreYou Shouldn't Need to Ask How to Make Your Agents Work Together
Matt Shumer asked how to run multiple OpenClaw agents together. The answer reveals a fundamental architecture gap — and why SpaceBot was built differently.
Read moreWhat I Learned Building Production Agents on SpaceBot
Hard-won lessons from running production AI agents on SpaceBot — scheduling gotchas, undocumented APIs, and the management dashboard I built because it didn't exist yet.
Read moreAI Agent Frameworks in 2026: A Builder's Honest Comparison
I've built production AI agents with SpaceBot across SEO automation and automated bug fixing. Here's what I've learned about choosing an agent framework.
Read more