fishaudio/fish-speech ? reverse-engineered prompt

I want a working local setup of Fish Speech so I can type text and generate really natural sounding speech, ideally in multiple languages, with optional voice cloning from a short reference clip. Please make it easy to use, with a simple web interface if that’s available, and also leave me a command line way to run it.

I’d like support for expressive tags in the text like whispering, excited, sad, pauses, and similar tone controls, plus the ability to save the generated audio files. If there’s a server mode, set that up too so it can be called from other apps later.

Please install and configure whatever this project needs, use Docker if that’s the smoothest path, download the right model, and make sure it actually runs end to end on this machine. Add a short README for me with exactly how to start it, how to generate speech, and how to do voice cloning. Also mention any license or usage restrictions I should know about. Look up the current docs online if you need to.