README: full setup (models by RAM, CORS per OS, in-app steps, LM Studio/llama.cpp, ElevenLabs key scopes, Render deploy + domain), controls, known gotchas table, live URL