### What feature would you like to see? When using Codex through CLI there are a few things to improve: 1. Indentation is too huge so it's hard to read. 2. Formatting like in GitHub or Claude Code would be a sweat bonus. ### Additional information _No response_

Potential duplicates detected. Please review them and close your issue if it is a duplicate. - #6752 - #6531 - #6850 - #6579 - #6505 *Powered by [Codex Action](https://github.com/openai/codex-action)*

Code change in CLI is hard to read

Resolved 💬 9 comments Opened Nov 26, 2025 by degisner Closed Apr 3, 2026

💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

What feature would you like to see?

When using Codex through CLI there are a few things to improve:

Indentation is too huge so it's hard to read.
Formatting like in GitHub or Claude Code would be a sweat bonus.

Additional information

_No response_

View original on GitHub ↗

9 Comments

github-actions[bot] contributor · 7 months ago

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

#6752
#6531
#6850
#6579
#6505

Powered by Codex Action

etraut-openai contributor · 7 months ago

What platform are you using? From the screen shot, it looks like MacOS.

What terminal emulator are you using? Are you seeing this issue with the standard MacOS terminal app?

It's not clear to me from your bug report whether the issue is related to model behavior (that is, the output of the LLM) or a terminal display issue. Please use the /feedback slash command to upload the session and post the thread ID here. That should give us additional details we need to investigate further.

degisner · 7 months ago

Thank you for taking a look! Here are more details:

MacOS Tahoe (v26.1)
Ghostty (v1.2.3)
The issue is related to the code change display formatting. It seems like there's 6-8 spaces indentation instead of usual 4 or 2.
I'm sorry, I'm not sure how to use the /feedback slask command.

etraut-openai contributor · 7 months ago

@degisner, to use the /feedback slash command, just type /feedback into the text input box, then select "bug" and upload your session. There's no need to file a new bug report, but please respond here with the resulting "thread ID". This will allow us to look up your session.

degisner · 7 months ago

@etraut-openai If I understand correctly, I need to use GitHub Copilot for this, right? I don’t have it. Is there any other way to do that? I have no clue.

etraut-openai contributor · 7 months ago

Just run codex resume and choose the session that corresponds to the screen shot you posted above. Then type /feedback in the text input box. It will walk you through uploading the session details and print a thread ID when it's done.

degisner · 7 months ago

Oh, right! Thank you for the help! Here's a thread ID: 019ac456-6ae8-7fa2-a48c-135d671ee413
It's not the same as in the screenshot above but with the same issue — code diff is hard to read.

vignesh07 · 6 months ago

Submitted a PR for this: https://github.com/openai/codex/pull/9637

etraut-openai contributor · 3 months ago

Closing stale bug report. We've made significant modifications to the TUI code since this was submitted. If you're seeing a similar issue, please open a new bug report.