mohitsoni48/TurboLLM ? reverse-engineered prompt

Build me a local LLM app that I can run with one command and use on my own machine.

I want a polished web interface for chatting with models, plus an API that works like OpenAI and Anthropic so tools like Claude Code can connect to it without changes. It should run fully offline, use the GPU as well as possible, and automatically pick sensible settings instead of making me guess a bunch of flags.

Make it work with local model files I already have, and let me point it at different local LLM engines if needed, including custom ones. It should show me whether a model will fit in VRAM before I load it, and give me live speed and basic hardware info while it runs.

I’m on Windows, macOS, or Linux, so keep it cross platform, lightweight, and don’t use Electron or Python. If you need to check current docs or compatibility details online, do that.