Continued interaction during processing times
Resolved 💬 8 comments Opened Nov 29, 2025 by bpmiranda3099 Closed Jan 31, 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?
Currently, when Codex is doing a big task (like processing a large request or generating complex code), the chat locked unless you stop the current process. I have to wait for it to finish completely before I can type or ask anything else.
Please allow the long process to continue running in the background, but keep the chat input active. This way, I can still ask quick questions or provide more details without waiting for the first task to stop.
Additional information
_No response_
8 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Are you using codex web, CLI, or the IDE extension?
You can create multiple concurrent tasks:
If you're using the CLI or extension, you can stop the task at any time and enter a new prompt that modifies or adds to the previous prompt. The model will pick up where it left off and continue working with the new instructions.
@etraut-openai Yes, but there is no functionality equivalent to "branch this chat here". I agree that that would probably have a clumsy naive implementation in the extension, but actually I think OP's suggestion would be better; have the extension present itself as though it is working in the background, while allowing "foreground chat" backed by simply branching the history?
@etraut-openai, I am using the IDE extension.
@khoek exactly. The goal is to avoid managing multiple windows or losing context whenever possible. Enabling a "foreground chat" to continue while a main task runs in the background would be a really helpful improvement.
@bpmiranda3099, in the extension, you can easily create multiple concurrent conversations and flip between them. There's even a control at the top of the extension panel that indicates how many conversations are active at the moment so you can monitor concurrent tasks. You can easily flip between conversations using this control. This mechanism also allows you to have a foreground conversation while codex is working on one or more background tasks. Have you tried this? I frequently use the extension in this manner. In particular, I find it useful to use the cmd+N shortcut (ctrl+N on Windows) to quickly create a new conversation while codex is working on a task.
@khoek, "branching" an existing conversation into a two conversations is an interesting idea, but I don't think that's what the OP is asking for here. If you'd like to suggest that feature and flesh out how it might work, it's probably best to open a new enhancement request.
@etraut-openai , Thanks for the response. However, creating a new conversation separates the context.
It isn't just to do something else while waiting; it is to interact with the current task while it processes. For example, if I realize I forgot a requirement or want to queue a follow-up question related to the code currently being generated, I want to do that in the same thread so the context is preserved. Opening a new chat forces me to restate the context. Once the codex finishes the heavy task, "merge" my new inputs back into the main conversation history.
If you want to restate or clarify the task, add a new requirement, or provide additional instructions in the context of the current task, you can do that at any point. Just enter a new prompt. Have you tried that approach?
This feature request has not received enough upvotes in the past two months, so I'm going to close it.