There are two broad ways to work with a coding agent today. One is a terminal tool such as Claude Code, Codex CLI, or Aider, where the agent reads and edits files on its own and you watch the output scroll by. The other is an editor with the agent built in, such as Cursor or VS Code with Copilot, where edits show up inline next to the code you are looking at.
Each setup nudges you toward a different way of working. The terminal makes it natural to hand over a whole task and review the result afterwards. The editor makes it natural to stay in the loop line by line. Plenty of people use both, but for different jobs, and the split is rarely the same from one person to the next.
We are curious about your split:
- Where do you run your agent most of the time, and for which tasks?
- What made you switch from one to the other, if you did?
- Is there a task you would never do in the setup you use less?
Tell us about the setup you actually use day to day in the comments below.