OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Chat & Assistants

Miko

Miko is a Windows desktop AI companion that connects an Ollama-backed assistant to a Unity 3D avatar, with speech synthesis and real-time lip-sync. It supports Fish Audio cloud streaming or local GPT-SoVITS voice generation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Miko — OpenViST project profile in Chat & Assistants.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Miko when you want an Ollama-connected desktop companion whose responses can be spoken and represented by an animated 3D avatar.

Who it is for

  • Windows users seeking a conversational AI companion presented as an animated desktop character.
  • Developers interested in adapting an open-source assistant that combines Ollama, voice synthesis and Unity avatar presentation.

Documented uses

  • Hold text or voice conversations with an Ollama-connected assistant.
  • Hear generated responses through Fish Audio streaming or local GPT-SoVITS synthesis.
  • Display speech with an animated character and synchronized lip movement.
  • Explore a codebase combining an assistant engine, speech services and a Unity desktop avatar.
Before you choose

Requirements & limits

The project is a public GitHub repository identified as GPL-3.0 licensed. Its documented target is 64-bit Windows 10 or 11.

  • Ollama and a locally pulled model are prerequisites for the assistant.
  • Fish Audio requires an API key and uses cloud streaming; GPT-SoVITS uses local model files and is described as offline.
  • The README recommends an NVIDIA CUDA GPU for local GPT-SoVITS inference; Fish Audio voice generation is described as requiring no local GPU VRAM.
  • The desktop avatar is a Unity application, while the assistant and voice integrations are separate components.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare Miko for a voice-engine option

Install the documented environment and configure either Fish Audio or local GPT-SoVITS prerequisites.

Before you start

  • 64-bit Windows 10 or 11
  • Python 3.10 or 3.11 with pip
  • Ollama installed and a model pulled, such as qwen2.5:7b
  • The Miko repository, including setup.bat and .env.example

Steps

  1. Run setup.bat from PowerShell or Command Prompt to create the virtual environment, install dependencies, prepare .env and run the repository's setup test suites.

  2. For Fish Audio, obtain an API key and set FISH_API_KEY and the documented Fish Audio settings in .env.

  3. For local GPT-SoVITS, set the configuration, checkpoint and reference-audio paths in .env to the local files.

Expected outcome

The documented Python environment is prepared and the selected speech engine's configuration values are set; the README describes separate launch options for each engine.

Guide limitations

  • An NVIDIA GPU is optional for Fish Audio voice generation and recommended for local GPT-SoVITS inference.
  • Fish Audio mode requires an API key; the source does not document account or service terms.
  • This is a setup recipe, not a claim that installation or inference was independently tested.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
8 October 2026
Original website title
danukim/Miko · Official homepage title · retained in its source language
How we review sourcesSuggest a correction