ZephInk

The Engine

What ZephInk currently runs on — swap any of these independently, nothing is locked together.

🧠

Model Hosting

  • Embedded — a small model downloads and runs in-process. No separate app to install.
  • Local — points at your own LM Studio or Ollama server, with any model you like.
  • Cloud API — OpenRouter, Groq, or OpenAI, your own key.
🎙️

Voice Engine

  • Chatterbox (MIT-licensed) — reference-clip cloning, no license restrictions on commercial use.
  • Or point voice generation at a custom endpoint — your own self-hosted server, or a cloud API like ElevenLabs.
🖥️

Platform

  • Packaged Windows desktop app — no browser tab required.
  • Runs fully offline once its models are downloaded.
  • Automatic updates, with an opt-in Test/Beta channel.
v2.1.29 Last updated Oct 31, 2026 Created by Samuraikid11