feat: verbose output mode
Resolved 💬 5 comments Opened Apr 18, 2025 by hitachi-development Closed Aug 7, 2025
💡 Likely answer: A maintainer (tibo-openai, collaborator)
responded on this thread — see the highlighted reply below.
Codex currently displays a couple line summary of the output of commands it runs. But I find this really harms my ability to work with the AI well. If commands are encountering problems, the useful information is rarely in the first few lines. And by being able to see them, I can understand the context and aid the AI in troubleshooting. When its hidden I can't help out.
Until this is improved I find codename goose far more viable for real world tasks.
5 Comments
What additional information should codex output that would help you be more productive?
I guess there should be a command to expand all the output of the last step that is currently being truncated. There should probably be a toggle for verbose mode too.
I know there's the "explain command", but that's always not that important, all we need is to see the full picture of what just happened. Just adding "show full output", avoiding LLM roundtrip, would suffice.
Thank you for the feedback! Please try the latest version of Codex CLI to see if this issue persists. We're actively working on improvements and would appreciate your input on the native/Rust implementation of Codex CLI.
Not really. The goal is to be able to see a clearly understandable diff of the changes that are being done when changes occur. Amazon q chat does this very well. It allows people to have far greater confidence that the AI is doing the right things.
Without improvements here I have to let the AI do the changes then do git diffs myself which is really awkward.