Loading…
Turn 2D images into game-ready 3D assets instantly. Jupetar runs 100% locally on your NVIDIA GPU—no subscriptions, no cloud queues, no data leaks. Just drag, drop, and export high-fidelity GLB models for Unreal, Unity, and Blender. You own the software; you own the assets.
No offers tracked in PLN yet.
TURN 2D IMAGES INTO 3D BASE MESHES IN MINUTES. Jupetar is a powerful, locally-hosted generative 3D pipeline. Built on the cutting-edge Hunyuan3D architecture, it transforms a single 2D image—a concept art sketch, a photo, or a texture—into a fully textured GLB model to kickstart your 3D workflow.
Forget monthly subscriptions, cloud queues, and data privacy risks. Developed by Odyssey Game Studios, Jupetar runs 100% locally on your hardware. You own the software, you own the models, and your unreleased assets never leave your machine.
EXPECTATION MANAGEMENT: WHAT JUPETAR IS (AND ISN'T) Jupetar is designed to be a massive workflow accelerator, not a magic bullet. The models generated are base meshes. While it saves hours of manual block-out time, the raw GLB exports are not strictly "game-ready" out of the box. You will encounter micro-transgressions, minor artificing, or topology that requires a quick cleanup pass in Blender before dropping it into your game engine.
CORE FEATURES
⚡ Titan-Class Generation Engine: Leverages the raw power of NVIDIA CUDA cores to generate high-fidelity geometry and textures simultaneously. What used to take hours of manual block-outs now happens while you grab a coffee.
🔒 Zero-Cloud Architecture: Designed for privacy-conscious developers. Jupetar installs as a local, self-contained pipeline on your PC. No internet connection is required after the initial setup.
🎨 Intelligent Texture Synthesis: The engine doesn't just project the image; it understands the surface. It hallucinates missing details for the back and sides of your object, generating Albedo and Normal maps automatically.
🛠️ Workflow-Ready Output: Exports directly to the industry-standard .GLB (glTF) format. Ready to be imported straight into Blender for final topology cleanup, PBR refinement, and rigging before heading to Unreal Engine 5, Unity, or Godot.
DESIGNED FOR:
Indie Developers: Populate entire levels with background props and clutter in a single afternoon.
Modders: Turn 2D sprites from classic games into 3D objects for modern remakes.
Concept Artists: Rapidly prototype 3D block-outs from your 2D sketches to verify forms and lighting.
ROADMAP: FUTURE UPDATES Jupetar is actively evolving. We are currently working on the following post-launch implementations:
Enhanced Texturing: Major upgrades to facial details and overall texture resolution quality.
AMD GPU Compatibility: Active development is currently underway to expand our hardware support and bring the full Jupetar generation pipeline to AMD Radeon graphics cards.
Surface Area Review: Refining how the engine handles neighboring pixel textures prior to export (addressing blind spots so the bottoms of feet, shoes, and undersides are fully textured).
Engine Optimizations: Further under-the-hood optimizations for faster generation speeds.
Broader Compatibility: Ongoing adjustments to support a wider range of hardware architectures.
⚠️ STRICT SYSTEM REQUIREMENTS WARNING ⚠️ Jupetar is a highly demanding local AI application. Because it loads massive neural network weights directly into your graphics card, VRAM is an absolute requirement. Please review carefully before purchasing to ensure your system can handle the pipeline.
Minimum (Slower Generation): NVIDIA GPU with at least 10GB of VRAM (e.g., RTX 3060 12GB, RTX 2080 Ti 11GB, RTX 3080 10GB).
Recommended: NVIDIA GPU with 12GB to 16GB+ VRAM (e.g., RTX 4070, RTX 4080, RTX 3090, RTX 4090).
Not Supported: AMD Radeon (currently in development), Intel ARC, Integrated Graphics, or any NVIDIA GPU with 8GB of VRAM or less.
