The Data Engine for Clinical Physical AI
We capture credentialed OR, sterile-processing, and clinical workflow data to train surgical-robot foundation models — the only environment Mecka, Lightwheel, and Encord can't enter.
Available for macOS, Windows, and Linux
Built for robotics teams at
Everything you need to build robots
The complete robotics development pipeline — from natural language description to production-ready deployment. Every stage connected, every tool integrated.
Design
AI generates robot structure from natural language
Validate
Physics linter checks inertias, joints, collisions
Configure
ros2_control hardware interfaces & controllers
Program
Visual Behavior Trees with live XML sync
Simulate
Real-time physics preview with Rapier3D
Export
URDF, SDF, MJCF for any platform
Deploy
Production-ready ROS2 packages
Deep integration at every step
AI Robot Generation
Design → Validate
Describe your robot in natural language. Our AI generates complete kinematic chains, proper joint limits, and physically-valid inertial properties.
- Multi-model support (GPT-4, Claude, Gemini)
- Iterative refinement with chat
- Structured output validation
Physics Linter
Validate → Configure
Real-time physics validation catches errors before simulation. Warns about impossible inertias, self-collisions, and joint singularities.
- Inertia tensor validation
- Collision geometry checking
- Joint limit analysis
ros2_control Generator
Configure → Program
One-click generation of hardware interfaces, controller configs, and protocol-specific boilerplate for UART, CAN, I2C, and SPI.
- Hardware interface scaffolding
- Controller YAML generation
- Resource claim validation
Behavior Tree Editor
Program → Simulate
Visual drag-and-drop editor for BehaviorTree.CPP. Changes sync bi-directionally with XML — edit visually or in code.
- Live XML synchronization
- Node library with search
- Subtree composition
Physics Simulation
Simulate → Export
Instant physics preview with Rapier3D WebAssembly. Test kinematics, reachability, and collision detection without leaving the IDE.
- Real-time collision detection
- IK mode for reachability
- Gravity and friction testing
Multi-Format Export
Export → Deploy
Export to URDF for ROS, SDF for Gazebo, or MJCF for MuJoCo. Full mesh embedding and material preservation across formats.
- URDF, SDF, MJCF support
- Mesh embedding options
- ROS2 package scaffolding
The Kindly Ecosystem
Beyond the IDE — research infrastructure, enterprise solutions, and social impact initiatives
FoodforThought
Open data infrastructure for Physical AI. Community-powered labeling, demonstration capture, and full lineage tracking from sensor to skill.
KindlyforYou
Enterprise robotics platform. Fleet management, deployment orchestration, and scalable infrastructure for production robotics.
Learn More →KindlyforUs
501(c)(3) nonprofit. Community-driven robotics for social good — environmental cleanups, disaster response, and humanitarian aid.
Get Involved →