| Key features | - 40+ AI model lanes including Trellis, Hunyuan 3D, Tripo3D, Rodin, and Meshy
- Multi-view reference slots for front, back, left, and right images
- Downloadable textured GLB exports with task history
- 50+ free browser tools for inspection, conversion, and scaling
- 3D Model Viewer for GLB, OBJ, STL, and FBX formats
- 3D Viewer & Editor with local GLB, glTF, OBJ, PLY, and STL export
- GLB Inspector for structure, scenes, meshes, materials, and animations
- Texture Checker for GPU memory verification before export
- Image to STL conversion for 3D printing
| - First-person video capture with head, helmet or chest-mounted cameras
- Dual-reviewed annotation of tasks, tools, objects and outcomes
- Privacy processing including face and text removal before release
- Documented protocols agreed before recording begins
- Contributor consent tracking and payment per accepted hour
- Provenance records with capture context, protocol version and processing history
- Evaluation, research, commercial and enterprise licence tiers
- Custom dataset collection for specified tasks and environments
- Formats: MP4 video, Parquet tables, JSON manifest
- Immutable, versioned releases for reproducible experiments
|
|---|
| What makes it different | - No sign-up required for public queue tests and free browser tools
- 50+ local browser tools that never upload files or spend credits
- Preview and inspect outputs before paying with credits or downloading
- Supports 40+ AI models in one workspace without switching tools
- Clear pricing per lane before running, with credits valid for up to 1 year
| - Real-world field recordings from skilled professionals instead of lab or simulated data
- Includes failure and recovery episodes with explicit outcome state labels
- Contributors are paid upfront agreed rates with transparent review decisions
- Quality assurance with sample re-annotation and written acceptance criteria
- ISO 27001 and SOC 2 Type II certification in progress
|
|---|
Photo3D
Curatrix