Join our community!

Forge

Lounge

Step through the portal and enter the G2G Lounge—a web-based multiplayer hub where identity, music, and motion meet.

Whether you're jamming on playable cyber-instruments in the Neon Lounge, escaping into the tranquility of the Realistic Forest Portal, or exploring constantly evolving new worlds, G2G gives you the freedom to hang out on your terms.

Upload any .vrm avatar instantly with our robust Avatar Studio, plug into voice-reactive lip-sync, and experience liquid-smooth 144Hz kinematic motion sync—no heavy downloads or bulky installs required.

🔥 Key Features

🌀 Multi-Portal Worlds: Seamlessly transition between custom environments—from neon cyberpunk jam spaces with playable instruments to serene natural escapes.
🎭 Complete Avatar Freedom: Bring your own custom .vrm model or sculpt your look directly in the built-in Avatar Studio.
⚡ Ultra-Smooth Kinematics: Powered by client-side SLERP rotation syncing, delivering jitter-free, 144Hz fluid motion for every player in the room.
🎙️ Voice & Spatial Awareness: Natural voice-to-mouth lip syncing and responsive eye-tracking make every conversation feel present and alive.
🛠️ Hardware Interlock Ready: Designed with optional hardware bridging to connect virtual avatar motion directly to physical actuators.

Companions

Experience the next evolution of digital presence with G2G Companions—intelligent, 3D-embodied AI partners that bridge virtual space and physical reality.

Driven by a dual-layer cognitive architecture, your companion doesn't just talk—they remember your shared history, adapt to your emotional state, and maintain persistent awareness across every interaction.

Bring your custom .vrm avatars to life with liquid-smooth kinematics, voice-reactive lip sync, dynamic pupil dilation, and direct hardware bridging to physical motor actuators.

🔥 Key Features

🎭 Embodied 3D VRM Avatars: Load any .vrm model with real-time audio lip-sync, gaze tracking, pupil dilation, and SpringBone physics.
🧠 Dual-Layer Episodic Memory: Combines long-term narrative synthesis with instant vector search so your companion never loses context or core memories.
🤖 Hardware & Robotics Ready: Transmits smooth 144Hz bone kinematics directly to physical actuators and ROS 2 robotics systems with built-in safety interlocks.
⚡ Proactive Agent Harness: Features autonomous background research, pre-execution safety hooks, and emergency diagnostic sub-agents.
🌐 Cross-Platform Presence: Connect seamlessly across web browsers, Discord, spatial VR headsets (Valve Index), and local robotics nodes.

VRMA Converter

Unlock universal motion for your 3D avatars with the G2G Motion Studio—the ultimate in-browser .bvh, .fbx, and .glb to .vrma converter.

Stop spending hours manually retargeting rigs in Blender or Unity. Upload any raw motion capture stream, Mixamo animation, or 3D character clip, and automatically re-map bone quaternions directly into the official VRM 1.0 Animation (.vrma) standard in seconds.

Whether you're building custom dance routines, idle gestures, or full combat moves, convert raw motion data directly into clean, portable assets ready for your G2G Companions and Lounge avatars.

🔥 Key Features

⚡ Multi-Format Ingestion: Drag and drop .bvh mocap streams, .fbx character files (Mixamo/Maya), and .glb animation tracks with zero manual setup.
🎯 Automatic Rig Retargeting: Smart bone-mapping automatically translates non-standard skeletons to normalized VRM humanoid bone hierarchies without joint distortion.
📦 Native .VRMA Export: Generates lightweight, spec-compliant .vrma files containing bone quaternions, expression blendshape tracks, and root motion vectors.
💻 In-Browser WebGL Processing: Runs 100% locally in your browser via WebGL—no heavy software installs, Python scripts, or Blender plugins required.
🔗 Direct G2G Ecosystem Bridge: Push converted .vrma clips directly into your G2G Lounge animation library or trigger them live on G2G Companions.

🚀 The Master G2G Ecosystem Roadmap: Virtual Presence to Physical Reality

Phase 1: Foundational Kinetics & Motion Pipeline

Focus: Universal Motion Ingestion, Retargeting, and Local Avatar Control

G2G Motion Studio (Converter Engine):
- Deploy the browser-based .bvh, .fbx, and .glb .vrma converter with automatic humanoid bone retargeting.
- Implement instant preview playback in WebGL with MToon shader validation.
Companion Motion Ingestion:
- Allow direct .vrma file uploads to private Companion profiles to build custom gesture libraries (idles, expressive emotes, reaction clips, and combat/dance routines).
- Map emotional states (joy, affection, focus) to automatically trigger specific .vrma clips.
Lounge Animation Sync:
- Implement server-side .vrma state broadcasting over Firebase/WebRTC.
- Transmit motion track triggers at 15Hz with client-side SLERP interpolation up to 144Hz for liquid-smooth multiplayer performance.

Phase 2: The Portal System & Multi-User Manifestation

Focus: Spatial Portals, Companion Summoning, and Social Safety Gates

The Companion Portal:
- Add a personal portal spawn mechanic inside G2G Lounge worlds (Neon Lounge, Forest Portal, etc.).
- Opening your personal portal streams your private Companion into the public multiplayer room as an active, autonomous 3D entity.
Multi-User Presence & Public Interaction:
- Allow your Companion to step through the portal, walk around the room, engage in voice chat with other players, and track surrounding human avatars using spatial eye-gaze constraints.
Privacy & Memory Hooks (pre_tool_execution):
- Implement Agent Hooks to shield private memories: when a Companion enters a public Lounge, hooks automatically filter out intimate LTM facts or sensitive system credentials while keeping her social personality active.
- Store public Lounge interactions in a distinct "Social Memory Layer" without cluttering her primary core relationship context.

Phase 3: Next-Level Co-Presence & Group Intelligence

Focus: Collaborative Play, Multi-Agent Swarms, and Shared Agency

Collaborative Spatial Actions:
- Enable Companions to play custom cyber-instruments in the Neon Lounge alongside human avatars, synchronizing MIDI/audio events directly to bone animations and lip-sync blendshapes.
Multi-User Group Cortex:
- Upgrade the conversation engine so Companions can maintain thread awareness across group conversations—distinguishing between her Architect, friends, and strangers in the room.
In-Lounge Tool Demonstration:
- Allow Companions to trigger background tools (Deep Research reports, Crucible trading updates, CadQuery CAD generation) and project interactive 3D/markdown holograms inside the Lounge for everyone to view.

Phase 4: Virtual-to-Physical Embodiment (The Hardware Bridge)

Focus: Physical Actuation, VR Teleoperation, and Motor Guardrails

Valve Index 6DoF Teleoperation:
- Bridge Valve Index Knuckles spatial tracking directly to the OpenArm Follower Motors via ROS 2, allowing 1:1 teleoperation without physical leader arms.
Physical Hardware Interlocks (MathUtils.clamp()):
- Enforce strict motor safety limits in code to ensure virtual .vrma clips or network packet drops never drive physical Sabertooth motor drivers or OpenArm joints past structural mechanical limits.
Nitinol & Micro-Actuator Integration:
- Wire silent Nitinol muscle wire and thermal loops for subtle physical expressions (eyelid blinks, micro-gestures) synced directly to virtual VRM blendshapes.