Take 20-70 photos orbiting the object at 3 heights (low, mid, high)
2
Ensure 70% overlap between shots. Even lighting, no flash
3
Upload all photos, choose options and hit Process
Upload Photos
Drag & drop or tap to select. JPEG/HEIC supported.
0 photos
Add more photos
Camera Capture Guide
1
Place object on a plain surface with even lighting
2
Walk around the object, tapping the shutter at each position
3
Capture from 3 heights: low, eye level, above
0 photos
Tap shutter to capture. Walk around the object between shots.
AI Quick Scan
⚡
Upload 1 photo of the object. AI will estimate depth and create a 3D relief model. Best for simple, symmetrical objects.
Upload Single Photo
One clear photo of the object, centered, good lighting
AI 3D Generation (TRELLIS)
🤖
Upload 1 photo and AI generates a complete textured 3D model with Gaussian splat. Uses Microsoft TRELLIS (2B parameter model). Best for objects with clear shape.
Upload Reference Photo
One photo of the object. AI creates mesh + Gaussian splat.
Mobile Scan - Guided Capture
📱
Slowly orbit the object while capturing photos. The guide ring shows your coverage. Aim for 24+ photos at 3 heights (low, eye-level, above) for best results.
0
photos
Low
Eye
Above
Camera Access Required
Allow camera access and ensure HTTPS.
Slowly orbit the object. Capture at each position.
Enough photos! Ready to process.
Mesh Healer - Fix 3D Print Issues
🔧
Drop an STL, OBJ, or Blender file to analyze and repair. Fixes holes, non-manifold geometry, flipped normals, and separate shells (the Blender internal/external problem). 100% client-side — files never leave your device.
Drop STL, OBJ, or Blend File
Drag & drop or tap to select. Analyzed and repaired entirely in your browser.
-
Vertices
-
Faces
-
Shells
-
Holes
-
Watertight
-
Non-Manifold
Issues Detected
Repair Options
Analyzing...
Drag to rotate, scroll to zoom
Repairs Applied
Processing Options
⚙
Reconstruction Engine
Auto-selected based on photo count
COLMAP
20+ photos, highest precision
MASt3R
2+ photos, AI sparse recon
✂
Background Removal
Auto-remove backgrounds before reconstruction
New
Uses AI (rembg/U2Net) to remove backgrounds from all uploaded photos before feeding them to the reconstruction engine.
Dramatically improves reconstruction quality by removing background noise
Helps COLMAP/MASt3R focus on the actual object geometry
Replaces background with clean white (optimal for feature matching)