LyubomirT/intense-rp-next — reverse-engineered prompt

Build me a Python desktop app that lets me use LLM web chat sites through SillyTavern like they were a normal OpenAI compatible API.

I want a simple window with a Start button, status, port setting, provider choice, and basic settings. When I start it, it should open a real browser, let me log in to services like DeepSeek, GLM Chat, Moonshot, or QwenLM, then expose a local endpoint at something like http://127.0.0.1:7777/v1. SillyTavern or any OpenAI compatible client should be able to send chat requests there and get streaming replies back.

It should handle one request at a time safely, show useful errors when the web UI changes or login is needed, and keep settings saved between launches. Please make it practical for normal users, with clear setup instructions and a way to run from source. Use browser automation and a local API server, and look up current docs online if needed.