# IDE agent or terminal agent: which coding agent setup actually stuck for you? — retained record

> Retained reddit record from reddit:AI_Agents.

Canonical: https://highsignal.app/data/records/1507d941d9f55aa4

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:AI_Agents)  published 2026-10-27 13:59 ingested 2026-10-28 02:35

[original source ↗](https://www.reddit.com/r/AI_Agents/comments/1xgtaoc/ide_agent_or_terminal_agent_which_coding_agent/)

## retained text



```
I'm curious which coding agents people kept using after the honeymoon period, not which ones they tried once.

Full disclosure: I built a small free rankings site for AI agent tools. In its coding agents category, Cursor is first at 8.75, Claude Code is 8.4, and Windsurf is 8.05. Claude Code shipped 2.1.310 on October 23 and already followed it with 2.1.311 on October 26, and its score held at 8.4 through both. Aider, GitHub Copilot and Continue come after that. One thing that surprised me: Aider scores the highest on voice coding fit in the group, but the newest release I track for it is from August 2025, so its freshness score is low.

My rough take is that the split is less about which tool is "best" and more about where you work. IDE-first people seem to settle on Cursor or Windsurf, while terminal people and anyone handing off whole repo-wide tasks lean toward Claude Code or Aider.

So: which coding agent do you actually use day to day, IDE or terminal, and did you switch at some point? If anyone codes by voice, what's your setup? If you think my scores are off, tell me and I'll fix them.
```



document fetched 2026-10-28 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xgtaoc/ide_agent_or_terminal_agent_which_coding_agent
