AI Workflow Lab

This isn't a product. It's a lab. I'm actively experimenting with AI agents, multi-step reasoning chains, workflow automation, and the kind of systems that can actually do useful things without constant human hand-holding. Some of it works. Some of it spectacularly doesn't. The goal is to understand how production AI workflows are actually built — not the tutorial version, but the version that handles weird edge cases, rate limits, and the occasional completely unhinged model output.