AI in Whiteboard Mode
Generate diagrams, modify existing content, and ask questions about your whiteboard — all through natural language prompts. AI updates are merged directly into your live scene.
Opening the AI Assistant
Access the AI command panel in two ways:
Toolbar Button
Click the animated sparkle icon at the right end of the whiteboard toolbar. It has a distinctive rainbow gradient border.
Keyboard Shortcut
Press ⌘/Ctrl + K from anywhere in the editor to open the AI panel. Works in both flowchart and whiteboard modes.
Three AI Modes
The AI panel offers three modes, each shown as a tab. Select the mode that matches what you want to do, then choose a suggestion template or type a custom prompt.
Create
Generate new contentAI generates a complete whiteboard scene from a text prompt. Shapes, arrows, text, and layout are produced automatically — you refine the result afterward.
Quick-action templates
Diagrams
- Simple flowchart
- Decision tree
- Architecture diagram
Wireframes
- Login page wireframe
- Dashboard layout
- Mobile app screen
Mind Maps
- Project brainstorm
- Feature mapping
- Concept exploration
Planning
- Sprint board
- Product roadmap
- Timeline with milestones
Edit
Modify existing contentAI reads your current whiteboard scene and applies modifications based on your instruction. Changes are merged into the live canvas so you can review and refine.
Quick-action templates
Style
- Change all rectangles to blue
- Add labels to shapes
- Set roughness to 0 for clean lines
Structure
- Connect shapes with arrows
- Reorganize into a grid layout
- Add a new section below
Cleanup
- Remove all arrows
- Remove standalone text
- Simplify by removing duplicates
Ask
Analyze contentAI answers questions about your whiteboard — describing content, listing elements, identifying gaps, and explaining the flow. No modifications are made to the scene.
Quick-action templates
Overview
- What's on this whiteboard?
- List all shapes and labels
- Describe the connections
Analyze
- Any missing pieces?
- What flow does this show?
- How are elements organized?
Image Attachments
Attach images to your AI prompts for visual context. This is especially useful for Create mode when you want the AI to reference an existing design.
5
Max images per prompt
5 MB
Max size per image
PNG/JPG/WebP
Supported formats
How AI Results Are Applied
AI generates elements
The AI returns a structured JSON response with shapes, arrows, text, and layout positions.
Elements merge into scene
For Create, new elements are added to the canvas. For Edit, existing elements are modified or removed. The scene updates live.
You refine manually
After AI applies changes, use the drawing tools to adjust positions, colors, text, and connections. Undo (⌘/Ctrl+Z) if the result isn't what you expected.
Tips for Effective Prompts
- Be specific about layout — "Draw a flowchart with 5 steps arranged top to bottom" works better than "make a flowchart."
- Name the elements — "Create boxes for Frontend, API Gateway, Backend, and Database" gives the AI concrete labels to use.
- Describe connections — "Connect Frontend to API Gateway with an arrow labeled 'REST'" tells the AI exactly what to draw.
- Use Edit for refinement — Generate first, then use Edit mode prompts like "make it cleaner" or "change colors to blue" for iterative polish.
- Attach reference images — When you have an existing design to replicate, attach it as an image so the AI can use it as context.
Tip: AI in whiteboard mode reads your entire scene as context. The more content you already have on the canvas, the more targeted the AI's edits and suggestions will be.