Robotics UI kit
The SYNCROBOTIC Design System is a UI kit for the people who run robots: React components for robot control apps, fleet consoles and teleoperation, built on shadcn/ui and Tailwind CSS v4. Each one installs into your project as source.
Most UI kits stop at buttons, forms and charts. A robot app also needs an emergency stop that is never hidden, a drive control that fails safe, a map in metres, a camera that says when it froze, and alarms an operator can trust at 3 a.m. Those are here, with the same tokens, dark and light themes, a high-contrast mode for sunlight and glove-sized targets for the field.
Safety
Stopping comes first: the stop is always on screen, and a dangerous action needs a deliberate gesture.
- Emergency stopThe emergency stop: one touch stops, holding 2 s stops all; lying down it becomes wake and stand up.
- Slide to releaseReleases a stopped robot with a deliberate slide: past 85% it releases, short of that it springs back.
- AlarmAlarms after ISA-18.2: three priorities, each with its own shape, colour and words, acknowledge and clear states, and a sorted list that announces new alarms.
Driving and missions
Manual driving that stops the robot the moment the finger lifts, and a map to send it somewhere.
What the robot sees
The camera, the LiDAR and the map it built, each saying in words whether the picture is live.
- Camera feedA robot's live camera over a native video the app feeds, that always says whether the picture is live: Live with the latency, Stalled with the last frame's time, Paused, connecting, offline or no camera.
- Point cloudA LiDAR or depth point cloud in 3D, drawn with WebGL2: coloured by height or intensity on one hue, over a ground grid, with the robot at the origin; turned, panned and zoomed by pointer, touch and keyboard.
- Annotated imageA picture with boxes, points, polygons or poses drawn over it: what a model predicted, what a person labelled, the previous version.
The robot's state
Pose, joints, battery and status, in shapes and words rather than colour alone.
- Robot modelA robot's pose in 3D from its URDF and live joint values, for monitoring: neutral greys, a ground grid, axes named in words, and the joints that need someone highlighted and named.
- Joint stateA robot arm's joints for monitoring: each joint's position in degrees or millimetres on a thin track between its URDF limits, with velocity and effort, and words when a joint is near, at or past a limit.
- Status glyphStatus is a shape and a word; colour only when it is abnormal. The shapes: filled circle busy, hollow circle normal, triangle attention, square stopped, slashed circle offline.
- MeterA level within a known range, like battery or storage, neutral until it crosses a threshold.
- Metric cardOne headline number on a card: what it counts, the value, how it changed and a line of context, coloured only when the change is for the worse.
App shell for robot apps
The frame of a control app held outdoors: a dynamic island for the one thing that matters now, a dock and trays.
- Dynamic islandThe top of the control app: name, device status and notifications, each expanding on its own, with the most urgent alarm inside.
- DockThe bottom bar of the control app in three tiers: the state action, this page's primary action, navigation.
- TrayThe shared frame of the island and the dock: radius 24, padding 8, gap 8.
- Grouped listRows on one surface with inset dividers; status as shape plus word.
Works with ROS and ROS 2
No ROS library inside: the components take plain props in the shape the messages arrive in (REP 103 axes, metres and radians), so whatever bridge your app uses passes them straight through.
| Message | Component | Props |
|---|---|---|
| nav_msgs/OccupancyGrid | robot-map | grid, costmap |
| sensor_msgs/LaserScan | robot-map | scans |
| geometry_msgs/PoseStamped (goal) | robot-map | mode="goal", onGoal |
| sensor_msgs/PointCloud2 (decoded) | point-cloud | points, intensity |
| sensor_msgs/JointState | joint-state, robot-model | joints |
| URDF (robot_description) | robot-model | urdf, urdfUrl |
| geometry_msgs/Twist | teleop-pad | onCommand: { linear, angular } |
| Video (WebRTC or a URL) | camera-feed | stream, src |
To get them there from a robot, use-ros connects over rosbridge with no ROS library and converts each message on the way: see Connect to ROS, with a live example and the safety notes for driving over a browser.
Install
Add the registry once (see Installation), then any component by name. robot-model is the only one that brings three.js; the rest add no 3D or map library.
npx shadcn add @syncrobotic/stop-button @syncrobotic/teleop-pad @syncrobotic/robot-map @syncrobotic/camera-feed