Hi everyone,
I'm trying to convince my small team (four of us devs) to use an AI coding assistant. We're all on individual GitHub Copilot plans right now, but I keep hearing about Continue and how it can be self-hosted or used locally. The "free" part is obviously super appealing to my manager.
But I'm a bit overwhelmed trying to figure out the *real* cost comparison. GitHub Copilot Business is $19/user/month. That's simple.
For Continue, the open-source version is free, but then I read about needing to provide your own API keys for models (like from OpenAI or Anthropic). That seems like it could get complicated and unpredictable.
So my main questions are:
- Has anyone actually tracked their monthly API costs using Continue with, say, GPT-4 or Claude, for a small team? Does it end up being more or less than the flat Copilot fee?
- Are there hidden infrastructure costs if we self-host the server? We'd probably run it on a cloud VM, so that's another line item, right?
- What about the time and effort to set it up and maintain it? That feels like a real cost too, especially since Copilot just works in the IDE.
I'm worried about proposing the "free" option, only for us to get hit with a huge API bill or spend a week debugging something. I just want a straightforward comparison for a small, bootstrapped team.
Any real-world numbers or experiences would be a huge help!