I don’t want to babysit my AI coding agents anymore
retained text
🦄 I’m the co-founder of a small AI company, and my co-founder Arseni is a software engineer. We use Claude Code and Codex constantly while building our products. One thing kept bothering us: the better these agents became, the more invisible management accumulated around them. Which one should take this task? Who has the context? Is the session still alive? Did it hit quota? Do I need to restart it? How do I carry the work from Codex to Claude without basically becoming human middleware? And honestly, I don’t want managing AI agents to become another job. That led us to an idea we’d been circling for a couple of years: **Claude Code or Codex doesn’t have to be the coworker.** The repo can have one persistent resident – identity, working memory, project context and open work – while Claude/Codex are replaceable execution engines underneath it. The process can die. The resident shouldn’t. We ended up building this as **brnrd**. One of my favourite uses is embarrassingly ordinary: I’m walking the dog, something about the product occurs to me, I message the resident from my phone and it starts working on our machine while I keep walking. We open-sourced it because something living on your machine and working with your repos shouldn’t be a black box. It has dogfooded itself pretty hard too: 1,000+ merged PRs in its own repo so far. **Disclosure because Reddit understandably hates disguised promotion:** yes, this is our project. I’m posting because I’m genuinely curious how people here solve the same problem 😇 If you use Claude + Codex (or several coding agents), what is your continuity layer today? Markdown files? tmux? Copy/paste? Your own scripts? An orchestrator? Repo, if you want to inspect/break what we built: [https://github.com/hugimuni-labs/brnrd](https://github.com/hugimuni-labs/brnrd)
document fetched 2026-09-18 00:41 · canonical https://reddit.com/r/ClaudeCode/comments/1wiw4d2/i_dont_want_to_babysit_my_ai_coding_agents_anymore