How I stopped context-switching between Claude Code, Codex, and GLM

For the past few months I've been juggling three AI coding agents — Claude Code for architecture, Codex for implementation, GLM for review. The pain wasn't any single one of them. It was the switching cost.

Every time I switched engines I had to re-explain the project: which framework, which conventions, where the old conversation left off. None of them remembered what the others knew. I'd waste 10–15 minutes per switch re-loading context, and half the time I'd just give up and stay on whatever I opened first.

So I started building a small thing for myself. Local desktop app. Lets me run all three engines from one window, with a shared memory layer underneath.

The thing I underestimated: persistent memory is the whole game. Once Claude and Codex can both see "this project uses SwiftData + strict concurrency + @Model on iOS 18+," switching becomes free. I stopped caring which engine I was on and started picking per-task.

Still rough around the edges. Town View (isometric map of conversations as villagers) sounds like a gimmick but it actually changed how I think about parallel runs. MCP bridge for moving work between machines was an accidental unlock.

Few things I'd love to hear from people actually doing this:

• For those running multiple agents — is the memory/switching problem real, or am I solving a me-specific issue?

• Has anyone found a workflow where the "let three models race, judge the output" thing actually beats just picking one?

• Anyone tried making their agents share persistent state? What broke?

Happy to go deep on the memory architecture if anyone's curious. Not selling anything — just trying to figure out if this is a real itch or a me-thing.