Annotation and Management of Embodied Intelligence Data
A one-stop platform for embodied intelligence data annotation and management, efficiently driving the implementation of large model applications.
One-click Export of Model Training Data
Human Data
Robot Data
Multi-channel Images, What You See Is What You Get
Depth Point Cloud, Perceive the Surroundings
Joint Pose, Record Millisecond-level Movements
Includes images, annotated natural language, joint pose changes, etc., to quickly start your model
Data Security is Paramount
The platform stores only data access links, not the data itself, and can quickly connect to various public/private cloud storage.
Protect your data privacy with IP whitelisting, access keys, role-based permissions, and data encryption.
- 腾讯云
- 阿里云
- 华为云
- Amazon Web Services (AWS)
- Microsoft Azure
- Google Cloud
- 自建私有云 MinIO
- 更多定制化支持...
Supports Multiple Data Formats
Supports playback and annotation of various formats, from simple images and MP4 videos to ROS (Robot Operating System) multimodal data Rosbag/Mcap.
- png
- jpg
- mp4
- webm
- bag
- mcap
- parquet
- tfrecord
Cross-Verification for Quality Assurance
Data undergoes multiple reviews by AI, reviewers, and project managers after initial annotation to ensure quality.
Developed based on extensive operational experience, ensuring separation between annotators and reviewers to maintain unbiased data review and performance evaluation.
High Performance, High Efficiency
Provides a smooth user experience, even on low-end computers and tablets, for multi-angle 3D robot simulations.
User-friendly interface with various shortcuts for quick learning and easy use.