endolith/open-interpreter ? reverse-engineered prompt

Build me a Python app called Open Interpreter that lets me talk to an AI in the terminal and have it run local code and shell commands for me after I approve them.

I want a simple command line chat experience where I can type plain English, see the assistant suggest actions, and then choose to run, reject, or edit them before anything executes. It should work as a local tool on my computer, with support for common things like Python and shell commands, and it should be usable both from the terminal and from Python code as a library. If it makes sense, include a few examples and basic docs so someone can try it quickly. Please make it feel polished and practical, with sensible defaults, clear safety prompts before execution, and a clean project structure with tests. If you need to check current docs or best practices for any model or API setup, look them up online first.