Real-time Avatar API
Runs directly in any web browser. Streams ultra-light motion packets (<1KB) via WebRTC in real-time. Supports multiple 3D avatar formats (VRM, GLB, FBX). On-device SDK coming soon.
Ask AI staff about the GOODGANG LABS.
They will answer everything.
Ultra-light WebRTC API streams real-time facial and body animation with zero device load—cross-platform, high-performance, and bandwidth-efficient.
Runs directly in any web browser. Streams ultra-light motion packets (<1KB) via WebRTC in real-time. Supports multiple 3D avatar formats (VRM, GLB, FBX). On-device SDK coming soon.
Cloud-scale engine for STT, LLM/SLM, TTS, motion(facial expression, body motion, lip-sync), and 3D Avatars. Modules are independent and replaceable — use our fine-tuned serving or swap in models like OpenAI, Gemini, Claude, or Grok. Seamless plug-and-play integration.
Create and serve fully interactive 3D AI Avatars in minutes. Launch instantly in the browser with a QR scan, no apps or hardware needed. Set personality and voice, upload text or PDFs, and auto-generate FAQs. Multilingual support included.
Flagship store in Seoul Gangnam. Voice-based consultation and ordering with 3D AI Avatars are already commercialized and running in-store. Partnering with diverse brands and pop-up experiences powered by our AI Avatar platform.
AI Engineer
Front-end Engineer
Back-end Engineer
AI Engineer
Product Designer
Administrative Manager
Backed by leading global investors and partners — plus individual angels from
Meta, Snap, Kakao, and other tech leaders.
GoodGang Labs is an AI avatar startup with a mission to connect people and AI through avatars. We are redefining human-AI communication by enabling natural, interactive experiences.
GoodGang Labs offers four flagship products: AI Staff, a voice-interactive avatar assistant; Marionette API, a real-time motion generation SaaS platform; KiKi Town, a 3D metaverse platform; and Beergang NFT, a 3D avatar collection that has generated over 1 billion KRW in sales. Each product leverages our avatar technology to create new digital experiences.
Our AI seamlessly integrates five technologies operating at 0.1-second precision: real-time STT, small language models, TTS with lip-sync, function calls, and voice-to-motion. Delivering 99% accuracy with 1.5-second response times, it creates natural human-like conversations in Korean, English, Chinese, and Japanese.”
The world's first AI avatar retail space.Experience shopping and checkout through natural voice conversations at our physical store located at 47 Gangnam - daero 160-gil, Sinsa- dong, Gangnam - gu, Seoul(Serosugil).
We've partnered with companies like Naver, Mamissson, and 78UNDER to deliver proven marketing impact through AI avatar- integrated flash sales and voice- interactive shopping.