Claude Code가 내 개발 환경에 대한 생각을 바꿔놓았다
Claude Code changed how I think about dev workspaces
핵심 요약
Claude Code 사용 시 늘어나는 작업 상태를 관리하기 위한 캔버스 워크스페이스 도구 'Cate'를 소개합니다.
- 개발 환경의 변화 — 터미널 하나로는 Claude Code와 연관된 여러 작업 상태를 관리하기 부족함
- Cate 도구 개발 — 터미널, 에디터, 브라우저 등을 통합 관리하는 캔버스 워크스페이스 오픈소스 프로젝트
- 워크스페이스 관리 — Git worktrees와 에이전트 작업을 결합해 복잡한 개발 환경을 시각화함
- 성능 최적화 — Electron 기반의 캔버스 환경에서 고부하 상황을 견디기 위한 최적화 기술 적용
I’ve been using Claude Code more as part of normal coding sessions, and it made me rethink something pretty basic: the terminal is starting to feel too small for the kind of work these tools do.
Not because Claude Code is bad in the terminal. It actually works well there. But the session around it grows quickly.
You have Claude working through changes, a dev server running, logs somewhere else, maybe docs open, maybe a browser preview, maybe a second branch or worktree because you don’t fully trust the first path yet.
At that point the problem is not only “what should I ask Claude?”
It becomes: where does this whole working state live?
I’m working on an open-source project around this idea called Cate. It’s basically a canvas workspace for terminals, editors, browser previews, and longer coding sessions. Not meant to replace Claude Code, more like a different surface around it.
Free to use, open source:
https://github.com/0-AI-UG/cate
Curious how others here are handling this.
Do you mostly keep Claude Code in one terminal, or are you already using split panes, tmux, multiple windows, worktrees, or several Claude sessions in parallel?
