Incorporate GPT 5 Prompt Optimizer into CLI with /optimize
Resolved 💬 5 comments Opened Aug 26, 2025 by mihir-s-05 Closed Jan 14, 2026
💡 Likely answer: A maintainer (etraut-openai, contributor)
responded on this thread — see the highlighted reply below.
What feature would you like to see?
Add the GPT 5 Prompt Optimizer within Codex to automatically optimize the prompt before sending it to get the best possible output. If it's possible to run the optimizer locally that'd be great; otherwise if it could just call the prompt optimizer and use it remotely that'd be great too.
Are you interested in implementing this feature?
If it's possible for me to get access to the prompt optimizer api, then yeah
Additional information
_No response_
5 Comments
this feature is actually available in the codex cloud when you press ask then it prepare the task for you to start with editable prompt. hope they add in the cli
/optimizeindeed seems like a logical addition! Could also see it coming to other CLIs.Big want for this, been using the optimizer recently and love the results. Just a little annoying going back and forth between a browser tab
This feature request hasn't received many upvotes in many months, so I'm going to close it.
Sharing a related external workaround since this issue was closed as not planned.
I built SuperGoal, a tiny open-source macOS plugin for Codex goal mode:
https://github.com/fightheyyy/SuperGoal
The idea is similar to a lightweight /optimize step: select a rough request, press one shortcut, and rewrite it into a clearer goal-mode prompt before sending it to Codex. The output is structured around objective, context, scope, non-goals, stop condition, acceptance criteria, and verification.
It supports a configurable OpenAI-compatible Base URL. With a cliproxyapi setup, it can use the same existing Codex route instead of adding another paid LLM provider/API key.
Not an official OpenAI feature, just sharing because this exact pain point is what pushed me to build it.