Getting started

The fast path from "I cloned the repo" (or "I downloaded the binary") to your first message in a chat.

1. Install — pre-built binary, Homebrew, or from source.
2. First run — the three-screen wizard: provider → model → display name.
3. Your first chat — built-in Pluma character, sending a message, switching characters.

Got a local LLM server already running (Ollama on 11434, LM Studio on 1234, mlx_lm on 1234, llama.cpp on 8080)? Pluma's auto-detect finds it. Don't have one? The wizard points at the model browser so you can download something.

Want HTTPS access from your phone? Do the chat-from-laptop loop first, then come back for Multi-device access (Tailscale).