Mintlify

The move

Helicone joined Mintlify in March 2026. I now build the context layer for AI agents at Mintlify. If you want the short version, the announcement covers it. This page is the longer one.

Why

In a world full of AI agents, the thing companies most need to invest in is a knowledge and context infrastructure layer. The analogy I used in the announcement: Waymo didn't reinvent the roads, but it still needed to learn about them. Agents are in the same position. They inherit the systems we already have, and they need a reliable way to understand them. Mintlify is the information layer for agents building the systems of the future. The overlap with Helicone was obvious to me from the first conversation: same world, different angle.

There was a simpler reason too. I genuinely like the team and the product. After talking to a dozen companies, that combination turned out to be rarer than you'd think. Cole and I both admire what Han and Hahnbee have built and how they operate.

What Helicone taught me

Three years of building Helicone left me with a few underrated lessons. The biggest one: a 5-person team can do an absurd amount if everyone is locked in. Headcount was never the constraint for us. Focus was. Mintlify runs the same way, which made the decision easier. You can read the full Helicone story here.

The context layer

Practically, the work comes down to one observation: agents are only as good as the context you hand them. You can swap models and tune prompts all day, and an agent with stale or missing context will still fail. Mintlify already powers documentation for thousands of dev tools, which is exactly the surface agents read when they try to use those tools. My job is to make that surface work as well for agents as it does for people. That is the context layer, and it is the most leveraged place I could be working right now.