Skip to main content
The CVAT annotation editor is the primary interface for creating and managing annotations. It provides a powerful set of tools for manual annotation, auto-annotation, and quality control across images, videos, and 3D point clouds.

​
Workspace Layouts

CVAT provides multiple specialized workspace layouts optimized for different annotation workflows:

Standard Workspace

Default annotation interface with full controls for creating shapes, tracks, and tags

Attribute Annotation

Focused interface for rapidly annotating object attributes

Tag Annotation

Streamlined workspace for frame-level tag annotation

Review Workspace

Quality control interface with issue tracking and commenting

3D Workspace

Specialized interface for annotating 3D point cloud data

​
Interface Components

​
Canvas Area

The central canvas displays your images, video frames, or 3D point clouds. The canvas supports:
  • Pan and Zoom: Navigate large images with mouse controls
  • Grid Overlay: Display alignment grid for precise annotation
  • Image Adjustments: Brightness, contrast, saturation, and gamma controls
  • Rotation: Rotate canvas for easier annotation of rotated objects
Use Ctrl+Wheel to zoom in/out, or hold Space and drag to pan around the canvas.

​
Controls Side Bar

Located on the left side, the controls sidebar provides access to annotation tools: Navigation Tools
  • Cursor tool (Esc) - Select and manipulate objects
  • Move tool - Pan the canvas
  • Rotate controls (Ctrl+R / Ctrl+Shift+R) - Rotate canvas clockwise/counterclockwise
Drawing Tools
  • Rectangle (N) - Draw bounding boxes
  • Polygon (N) - Draw polygonal regions
  • Polyline (N) - Draw open polylines
  • Points (N) - Place point markers
  • Ellipse (N) - Draw elliptical shapes
  • Cuboid (N) - Draw 3D cuboids (2D projection)
  • Mask (N) - Paint pixel-accurate masks with brush tools
  • Skeleton (N) - Annotate articulated objects with keypoints
  • Tag - Add frame-level classification tags
Advanced Tools
  • AI Tools - Interactive segmentation with SAM2 and other models
  • OpenCV Tools - Semi-automatic annotation helpers
  • Merge (M) - Combine multiple shapes
  • Group (G) - Group related objects
  • Split (Alt+M) - Split tracks at current frame
  • Join - Connect separate tracks
  • Slice - Cut objects with a line

​
Objects Side Bar

The right sidebar displays all annotated objects in the current frame:
  • Object List: View all shapes, tracks, and tags
  • Z-Order Layers: Objects grouped by rendering order
  • Quick Actions: Lock, hide, change attributes
  • Search & Filter: Find specific objects by label, attribute, or ID
Objects are color-coded by label. The active object is highlighted with a thicker border.

​
Top Navigation Bar

The top bar provides:
  • Job Information: Task name, job ID, frame range
  • Save Button (Ctrl+S): Save annotations to server
  • Undo/Redo (Ctrl+Z / Ctrl+Shift+Z): Reverse or replay actions
  • Finish Job: Mark job as completed
  • Menu: Export, settings, shortcuts reference

​
Player Controls

For video and image sequence annotation:
  • Play/Pause (Space): Auto-advance through frames
  • Previous/Next Frame (F / D): Navigate one frame at a time
  • Previous/Next Filtered Frame (Left / Right): Skip to frames with/without annotations
  • First/Last Frame (~ / <tilde>): Jump to job boundaries
  • Go to Frame (G): Jump directly to a specific frame number
  • Frame Step Controls: Configure auto-play speed
F          - Previous frame
D          - Next frame
Space      - Play/pause
Left       - Previous frame with objects
Right      - Next frame with objects
Ctrl+Left  - Previous empty frame
Ctrl+Right - Next empty frame
G          - Go to specific frame

​
Canvas Controls

​
Zoom and Fit

  • Fit Image (F1): Scale image to fit the viewport
  • Zoom In/Out (Ctrl+Wheel or Ctrl++/Ctrl+-): Adjust magnification
  • Reset Zoom: Return to 100% scale

​
Image Adjustments

Access the image settings panel to adjust:
  • Brightness
  • Contrast
  • Saturation
  • Gamma correction
  • Filters (grayscale, blur)
Image adjustments are visual only and don’t affect the underlying data or exported annotations.

​
Grid Overlay

Enable a grid overlay to align annotations:
  1. Click the grid icon in the settings
  2. Adjust grid size and opacity
  3. Use grid lines as visual guides

​
Context Menus

Right-click on the canvas or objects to access context-sensitive menus: Canvas Context Menu
  • Create new shape at cursor position
  • Paste copied shape
  • Configure canvas settings
Object Context Menu
  • Copy (Ctrl+C), Cut (Ctrl+X), Remove (Del)
  • Change label or attributes
  • Create object URL
  • Switch to track/shape mode
  • Propagate to next frames

​
Status Bar

The bottom status bar shows:
  • Current frame number and total frames
  • Active tool or mode
  • Selected object information
  • Zoom level
  • Cursor coordinates (when hovering over canvas)

​
Customization

​
Settings Panel

Access comprehensive settings via the menu (F2): Workspace Settings
  • Auto-save interval
  • Default object colors (by label, instance, group)
  • Text font size and position
  • Show/hide interpolated tracks
  • Enable/disable automatic bordering
Player Settings
  • Frame step (for videos)
  • Canvas rotation
  • Reset zoom on frame change
  • Show deleted frames
Keyboard Shortcuts
  • View all shortcuts
  • Customize key bindings
  • Reset to defaults
Press F2 anytime to open settings, or F3 to view the complete keyboard shortcuts reference.

​
Multi-Monitor Support

CVAT’s interface can be adapted for multi-monitor setups:
  • Detach player controls to a secondary monitor
  • Resize sidebar panels
  • Full-screen canvas mode

​
Workspace Switcher

Switch between workspace layouts using the dropdown in the top navigation:
  1. Standard - Full-featured annotation interface
  2. Attribute Annotation - Focus on attribute labeling
  3. Tag Annotation - Rapid frame classification
  4. Review - Quality control and issue resolution
Your workspace preference is saved per job and persists between sessions.

​
Performance Tips

  • Use browser zoom (Ctrl+Wheel) instead of canvas zoom for very large images
  • Enable “Reset zoom” option to avoid loading full resolution unnecessarily
  • Close unnecessary browser tabs to free memory
  • Use filtered frame navigation to skip empty frames
  • Enable auto-save at regular intervals
  • Consider splitting very long videos into multiple jobs
  • Use object filters to hide irrelevant objects
  • Lock objects you’re not currently editing
  • Group related objects to reduce clutter

​
Next Steps

Manual Annotation

Learn annotation workflows and techniques

Shapes and Objects

Explore all annotation types

Auto-Annotation

Use AI models for automatic annotation

Advanced Tools

Master interpolation, tracks, and propagation

Build docs developers (and LLMs) love

Get started for freeTalk to us