Random characters in TUI
Resolved 💬 14 comments Opened Oct 8, 2025 by distributedstatemachine Closed Oct 31, 2025
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
codex-cli 0.45.0
Which model were you using?
all
What platform is your computer?
Linux 6.8.0-64-generic x86_64 x86_64
What steps can reproduce the bug?
Start Codex
<img width="1070" height="382" alt="Image" src="https://github.com/user-attachments/assets/8e943614-fb8e-4c0e-a7ff-7d657cb0e13d" />
What is the expected behavior?
start codex
no ansi print out
What do you see instead?
Ansi print outs on screen
Additional information
I have heard of a friend having this when ssh-ing into a remote server, which is similar to my set up. This worked fine before i upgraded
14 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
The issue persists on latest master 26f7c468.
The issue persists on latest 0.47
Same here in both linux & macOS .. on latest 0.47
Same issue
same issue
The issue still persists in version 0.48. Any updates from the dev team on this annoying issue would be appreciated. Is it really that hard to fix?
Persists in 0.49
I finally managed to fix this annoying issue by using tmux.
Start a tmux session:
tmux new -s codexYou detach (Ctrl+b then d) and log out. The process keeps running inside tmux even if your SSH is gone.
Later, you can reconnect with ssh and run:
tmux attach -t codexand you’re back in the same terminal
Same here on a remote Azure server - only using tmux helped.
@nornagon-openai @etraut-openai @joshka-oai Could the bug https://github.com/PowerShell/openssh-portable/pull/806#issue-3483834510 be the cause of this issue (in case of sshing from Windows to Linux)?
This is still a problem on v0.104.0 in macos iterm.
<img width="1506" height="199" alt="Image" src="https://github.com/user-attachments/assets/a80d003d-a7ee-4bd2-a83c-72839c8dc4c2" />
Codex has a lot of really strong points, but this bug really kills usability for me.
@lidj22, this bug was fixed a long time ago. If you're seeing a similar behavior, it's likely a new or different problem. Please open a new bug report and provide as many repro steps or other details as you can. Thanks!
@etraut-openai thanks for the response! I will report details on this issue when available.