Sparse-view SFM-free Gaussian Splatting in Seconds
Generate 3D character models from single images
Create 3D models from YouTube videos
Image to 3D with DPT + 3D Point Cloud
Generate 3D models from text descriptions
Generate and refine 3D models from text prompts
Generate 3D model from a single image
Create 3D models from images or video
Generate 3D models from images
3D model reconstruction from video
VGGT (CVPR 2025)
Text-to-3D and Image-to-3D Generation
Turn 2D images into 3D models
InstantSplat is an advanced AI-powered tool designed to convert 2D sketches and multiple images into 3D models with remarkable accuracy. Leveraging Sparse-view SFM-free Gaussian Splatting, it achieves fast and reliable results in just seconds. Perfect for artists, designers, and 3D enthusiasts, InstantSplat simplifies the process of creating detailed 3D models from 2D inputs.
• Rapid Processing: Generate 3D models in seconds, making it ideal for time-sensitive projects.
• Multiple Image Support: Create accurate models from several 2D images taken from different angles.
• High-Quality Output: Produces detailed and precise 3D models with smooth surfaces and clean geometry.
• User-Friendly Interface: Easy to use, even for users with limited 3D modeling experience.
• SFM-Free Technology: Eliminates the need for Structure from Motion (SfM) techniques, reducing complexity.
1. How many images do I need to create a 3D model?
You need at least three images taken from different angles for optimal results, but more images will generally produce a more accurate model.
2. What file formats does InstantSplat support?
InstantSplat supports standard 2D image formats like JPG, PNG, and BMP for input, and outputs 3D models in formats such as OBJ, STL, and PLY.
3. Can InstantSplat work with sketches or hand-drawn images?
Yes, InstantSplat is designed to work with both photographs and sketches. However, for best results, ensure your sketches are clear and have defined edges.