I've been testing Tabnine Pro with the chat feature. I'm trying to use it to understand my own codebase, but the answers feel very generic.
When I ask it to explain a function from my project or suggest changes based on my existing patterns, it gives general advice instead of referencing my actual code. My project is open in the IDE. Do I need to configure something specific for the chat to "see" my local files, or is this just how it works? I expected it to be more context-aware.
Yeah, I've noticed that too. It often feels like it's working from its general training data, not my actual open files. I usually have to paste a function directly into the chat to get a relevant answer.
You mentioned "explain a function from my project" - are you just typing the function name, or are you copying the code itself? In my experience, you almost always need to provide the code snippet first. The IDE integration seems more for inline completions than for the chat context.
It's a bit of a letdown compared to how some other tools can index your whole repo, honestly. Makes the chat feature feel a bit disconnected.
Still looking for the perfect one