Flexpilot for AI Coding

The good stuff

What really clicked was the inline chat feature. I was stuck debugging a React state issue, and instead of jumping between docs and Stack Overflow, I just typed “Why is this state not updating?” right in the editor. Flexpilot gave me a clear explanation and even suggested a fix using useReducer. That saved me at least 15 minutes. The voice input is also surprisingly useful when I’m coding late and don’t want to type - just mumble your question, and it listens. Not perfect, but close enough to make hands-free work kind of fun. Also, seeing token usage in real time? Nice touch. Feels like I’m actually in control of what’s being sent out.

The not-so-good

It’s not flawless. Sometimes the suggestions feel a little generic, especially when working with niche libraries or custom frameworks. You’ll get the occasional “Here’s a basic loop” when you’re asking for something specific. And yeah, it’s still early - the community contributions are cool, but the UI feels a bit barebones compared to polished tools. If you’re used to Copilot’s slick interface, Flexpilot might look a little… plain. Also, since it’s open source and the developer isn’t public, there’s zero official support. If something breaks, you’re on your own unless you dig into the GitHub repo. So if you need hand-holding, this probably isn’t for you.

Bottom line

If you’re looking for a free, open-source alternative to Copilot that gives you actual control over which AI model you use - whether it’s OpenAI, Anthropic, or Gemini - then Flexpilot for AI Coding is worth a try. It’s not going to replace Copilot overnight, but for casual or experimental use, it works well and doesn’t cost a dime. I’d recommend it to anyone who wants transparency and hates paying for AI features they barely use. Just don’t expect magic. But hey, it’s free - so why not give it a spin?