All tools
View more
Access
Free
0
Research & Data Checked Aug 6, 2026
Large Spatial Model (LSM)
Turn two unposed images into a semantic 3D research reconstruction.
Best for: 3D vision developers and researchers exploring view synthesis, scene semantics, and radiance-field reconstruction.
What you can do with it
- Reconstruct a scene from two images
- Explore semantic 3D view synthesis
- Prototype a digital-twin pipeline
- Render novel views from indoor scenes
- Study open 3D vision model code
Good to know
- Live official project page
- Official implementation available
- Semantic 3D reconstruction
- Novel-view synthesis
Limitations
- Requires local technical setup
- GPU and model storage requirements
- Not a hosted consumer tool
Other official links
Keep exploring
Similar tools
Turn a single image into a first 3D shape reconstruction for inspection or further modeling.
3D & Gaming
Build physical-AI systems that can understand scenes, simulate outcomes, and learn actions.
3D & Gaming