[rjmacarthy]

Hello everyone!

I wanted to share something exciting I've been working on. After exploring various local alternatives to Copilot, I found that many required logins, proxies, or just didn't meet expectations. This inspired me to create my own solution.

After 132 commits, I'm now using this tool in my role as a front-end software engineer at a prominent cryptocurrency exchange. It's been an invaluable asset in my daily work.

Here are some details about the hardware compatibility: I use an Nvidia 3090, and the performance with Ollama codellama-13b is impressively swift. It's a great relief not to face the rate limiting issues common with Copilot. Additionally, I'm utilising codellama34b-instruct for the chat functionality. Its performance is on par with GPT-4, surpassing GPT-3.5 in quality. My friend has tested it on an Apple M1 Pro with great success.

If you're curious or looking for a reliable Copilot alternative, feel free to check it out at the link provided. Your feedback would be greatly appreciated.

I'm also looking for help with contributions, collaborations so reach out if you're interested. More features and updates to come in 2024.

Cheers!