Skip to content

Build

Build agents using the visual canvas and AI copilot.

The canvas is where you design your agent. Drag nodes from the panel, connect them, and configure each step.

TypePurposeExamples
TriggerStarts the agentSchedule, Webhook, App event
ActionPerforms operationSend email, Create issue
AIProcesses with AISummarize, Classify, Generate
ControlControls flowCondition, Loop, Delay

See Nodes Overview for full reference.

Every agent needs exactly one trigger:

  • Schedule - Run on cron (hourly, daily, weekly)
  • Webhook - Receive HTTP requests (coming soon)
  • App Event - React to events from connected apps (coming soon)
  • Manual - Trigger via button or API

AI Agent nodes process data with LLMs. Write a prompt describing what you need - the AI handles the rest.

Enable tools (web search, code execution, integrations) to give the AI autonomous capabilities.

Drag from output handle (right side) to input handle (left side) to connect nodes. Data flows through connections automatically.

Each node’s output becomes available to downstream nodes via {{nodeName.output}} syntax.

Instead of building manually, describe what you want:

Every morning, check GitHub issues labeled 'urgent',
summarize them with AI, and post to Slack

The copilot generates the full agent structure. Refine with the canvas if needed.