Cascades
Cascades enable AI agents to work together. When one agent needs help from another, it triggers a cascade — passing work to the right specialist.
What Are Cascades?
A cascade is when one AI agent hands off work to another. This happens automatically when a task spans multiple areas of your app.
For example, when you ask for "a contact form that sends emails":
- 1Canvas AI creates the form UI
- 2Canvas AI cascades to Backend AI: "need an endpoint to handle form submissions and configure email"
- 3Backend AI creates the cloud function AND configures SES for email delivery
How It Works
"Add a contact form that sends emails"
Canvas AI
Creates form UIcascade
Backend AI
Creates cloud function + configures SESCommon Patterns
Form Submission
"Add a contact form that saves to database"
CanvasBackend
File Upload
"Add image upload to the profile page"
CanvasBackend
Data Dashboard
"Show a dashboard with order statistics"
CanvasBackendDatabase
AI Chatbot
"Add a customer support chatbot"
CanvasAgents
Why Cascades Matter
- ✓Automatic coordination — You don't need to think about which agent to use
- ✓Full-stack features — Complex features get built end-to-end
- ✓Specialized expertise — Each agent is optimized for its domain
- ✓Consistent results — Agents communicate in structured formats
Cascades Are Automatic
You never need to manually trigger a cascade. Just describe what you want to build, and the system figures out which agents need to be involved. The agents communicate behind the scenes to deliver your complete feature.