I've been giving this a try for the past day and thought I'd talk about my experience setting it up in IntelliJ's terminal.<p>First, if you're going to use this in the IntelliJ terminal, make sure to use the dark-mode flag. Honestly the IntelliJ terminal doesn't do the best at rendering Aider's output because IntelliJ's terminal doesn't really like things that update frequently. It still works pretty well and Aider has nicely formatted output in general.<p>Using llama3.1 8B locally is OK at best, and makes incorrect changes at worst (the 8B model isn't really recommended in Aider's documentation, anyway). I only have 8GB of vram, so I can't really run any better models locally. I gave gpt-4o a try and it works pretty well at editing code. I probably won't pay for OpenAI access forever, though. <a href="https://aider.chat/docs/llms.html" rel="nofollow">https://aider.chat/docs/llms.html</a> says that Gemini 1.5 Pro can be accessed for free and works decently, so maybe I'll give that a try next.<p>The git integration is useful, but I don't really like that Aider will commit to my current branch to make it easy to undo its changes. I like changes being easy to undo, but I don't really want to have to squash these changes before pushing. I disabled the auto commits, and now I just have to make sure that I don't let Aider run wild on any file with a bunch of changes on it.<p>Maybe the automatic commits work for others' workflow, but it won't work for mine. I'll keep playing with this because it seems really cool.