Companion Nav
Open-LLM-VTuber

Open-LLM-VTuber

Open Source
t41372 (Yi-Ting) & Community

Open-source AI VTuber companion with Live2D avatar, real-time voice interaction, fully offline — cross-platform desktop app.

Visit Official Website
Product Type
Software
Platform
macOSWindowsLinux
Languages
English / Chinese
Regions
Pricing
Free
Price
Free & Open Source
Free Trial
✓ Trial Available
Completely free and open source
Last updated 2026-06-10· Content established

Data may be delayed — always check the official website for the latest info.

Visit Official Website

About This Tool

Open-LLM-VTuber is an open-source AI companion desktop app featuring Live2D dynamic avatars and real-time voice conversations. All inference runs locally with zero internet dependency. Core pipeline: Microphone → ASR → LLM → TTS → Live2D Expression Sync. Every module is swappable via config file (supports Ollama, OpenAI, Claude, DeepSeek, and more). Standout features include voice interruption (acoustic echo cancellation), emotion tag mapping (LLM emotion tags drive Live2D expressions in real-time), visual perception (optional), and desktop pet mode (click-through transparent window). v1.x stable release available, v2.0 under active development. 10,000+ GitHub stars.

Background

Open-LLM-VTuber was created by Taiwanese developer t41372 (Yi-Ting) with the goal of building a fully offline, customizable AI virtual companion. The project exploded on GitHub Trending in late 2025, quickly reaching 10,000+ stars.

Technology

Uses a modular pipeline architecture where every component (ASR, LLM, TTS, Live2D rendering) can be independently swapped out. Features voice interruption (acoustic echo cancellation), emotion tag mapping, and visual perception (camera/screen capture). Cross-platform on Windows, macOS, and Linux.

Market

One of the most notable projects in the open-source AI companion space. Covered by dev.to and other tech publications. Its primary audience is developers and tech enthusiasts, filling the gap left by closed-source commercial products like Neuro-sama.

Best For

Tech users who want self-hosted offline AI companionsLive2D and VTuber culture enthusiastsPrivacy-conscious users avoiding cloud servicesDevelopers who want full control over AI behavior

Editorial Note

The most complete open-source AI VTuber framework — offline, Live2D, voice interruption as core strengths.

Strengths

  • + Fully open source and offline with strong privacy protection
  • + Modular architecture with swappable components
  • + Smooth voice interruption experience
  • + Active community with 10,000+ GitHub stars
  • + Cross-platform (Win/Mac/Linux)

Drawbacks

  • Requires technical expertise to configure
  • v1.0.0 lacks long-term memory
  • v2.0 under active development, API may change
  • Live2D sample models are non-commercial only
Value AssessmentExcellent value for tech users — completely free and open source, zero cost with local models. Non-technical users should consider one-click solutions like z-waif.

External Resources

All information on this site is for reference only. Data may be delayed — always verify on the official website. Listings are in no particular order.