What we’re building

Case studies in agent design, perspectives on MCP apps, and practical guides for getting work done.

Subscribe via RSS

MCP apps

Industry: mobile gaming

method · 11 min read

To solve or not to solve

First we had to teach a level-playing agent how to play before we could use it to design good levels.

The game was brutally hard to solve, and nothing off-the-shelf came close — so the solver was the first deliverable. We built it with a second agent that played the game in a loop and reported back to Codex, which rewrote the simulator between runs. Only then could a level be graded at all.

Read the fieldnote
case study · 12 min read

No rulebook, no tells

We created a level-design agent without a rulebook and asked the designer of the studio in question to spot the additions of the agent.

The studio never gave us their rules. We reverse-engineered them from 100 shipped levels, built an agent on the guesses, and asked their own level designer to pick the machine's work out of a lineup — he caught 14 of 20, and still called the levels nearly impossible to tell apart from human work. Five days later his detection was no better than guessing, and the agent was shipping at 90% acceptance.

Read the fieldnote

Guides

Give your agent its next job.

Connect Recurse to your agent, and then ask it to do something for you there.