[Q] Is it possible to use aider --apply with output from web frontends like chatgpt.com?

[NightMachinery]

o1-preview is cheaper on the subscription on chatgpt.com, and in general, I like the flexibility of working with raw LLMs. But applying edits from the web frontend to local files is a PITA. I often tell the LLM to "Output the whole code", but this causes the web UI to crash if the code is long. I am wondering if I could add some formatting preamble to my chat requests, and then use aider --apply to apply the changes?

I use git, so after the change is applied, I'll be able to review the changes and accept/reject parts of them. But I need an easy way to apply the changes on long files in the first place.

[paul-gauthier]

Thanks for trying aider and filing this issue.

Sorry, no. There's isn't a way to do this. But it's an interesting idea.