Inspired by Open Interpreter – Feedback on Code-Atlas

[ystemsrx]

Is your feature request related to a problem? Please describe.

I’ve been exploring Open Interpreter and I find its ability to process natural language and execute code incredibly innovative. Inspired by it, I developed a similar concept called Code-Atlas, which also allows natural language to control a computer. However, while my project is still in the early experimental stage, I am looking for feedback to improve it and possibly make it more effective in its functionality, particularly for resource-constrained devices.

Describe the solution you'd like

I would like to refine my project by receiving feedback on how to better align it with the goals of Open Interpreter. Specifically, I’d like to know how I can improve its design, features (such as multi-language code execution and Markdown rendering), and general functionality to make it more efficient and user-friendly. Code-Atlas is implemented using C++ for better performance on devices with limited resources, so I’m particularly interested in optimizing the C++ implementation and ensuring cross-platform compatibility.

Describe alternatives you've considered

I’ve considered adding more features such as additional programming language support and better integration with various operating systems. However, I haven’t implemented these yet because I want to first focus on optimizing the current version for lighter resource usage and stability.

Additional context

I am currently a college student and Code-Atlas is just in its experimental, early stages. I’m particularly interested in learning from projects like Open Interpreter and exploring potential ways to enhance its capabilities. Any suggestions or advice would be greatly appreciated!

[Kreijstal]

ah very nice, I was looking for this, is there a way to use ollama?

[ystemsrx]

Yes, it fully supports Ollama. You just need to modify the config file, use the serve feature provided by Ollama, and set the base_url to http://localhost:11434/api/chat. This allows you to interact with the LLM through Ollama while also enabling you to adjust various parameters on it.

[Kreijstal]

Yes, it fully supports Ollama. You just need to modify the config file, use the serve feature provided by Ollama, and set the base_url to http://localhost:11434/api/chat. This allows you to interact with the LLM through Ollama while also enabling you to adjust various parameters on it.

since you are using c++ you can use std::filesystem instead of win32 api for more portability, also consider using cmake, and curses for terminal stuff

[ystemsrx]

Thanks! I'll consider std::filesystem for better portability, and also look into CMake and curses. Really appreciate your support!