tts-server-proxy

Chatterbox-TTS-Server (Windows portable) — ★☆☆ Easy

Repo: https://github.com/devnen/Chatterbox-TTS-Server · License: MIT GUI: ✅ full Web UI · Cloning: ✅ reference audio · Italian: ✅ Multilingual engine Default port: 8004

Self-hosts Resemble AI’s open-source Chatterbox family (Original, Multilingual — 23 languages, Turbo) behind an OpenAI-compatible API with a modern Web UI, chunked long-text handling and audiobook generation.

Why this one first

The Windows Portable Mode is the lowest-friction path that exists today: the launcher embeds its own Python, so the whole folder is self-contained — copy it anywhere (even a USB stick) and double-click start.bat. No Python installation required.

Install (Windows 10/11, ~10 minutes)

  1. Download/clone the repo: git clone https://github.com/devnen/Chatterbox-TTS-Server (or download the release zip).
  2. Run start.bat → first-time setup: choose Portable Mode (selected by default).
  3. Open http://localhost:8004 → the Web UI.
  4. In the UI: pick the Multilingual engine (required for Italian — do not use the Original engine for Italian), select a built-in voice or upload a short reference clip for cloning.
  5. Models download on first run (a few GB; NVIDIA GPU recommended — CPU works but is slower). RTX 30/40/50 supported.

Connect tts-server-proxy

App field Value
Endpoint http://<pc-ip>:8004
API key empty (no auth by default — keep it LAN-only)
model anything (server uses the engine selected in its UI), e.g. chatterbox
voice a built-in voice name or your reference clip name (see GET /v1/audio/voices)
lang it

Gotchas