Flow Builder
The Flow Builder lets you create visual workflows by connecting nodes and edges, automating multi-step processes that combine AI agent actions with integrations.
Overview
Workflows are directed graphs where:
- Nodes represent actions (AI tasks, API calls, data transformations)
- Edges define the execution order and data flow between nodes
Creating a Workflow
- Open the Flow Builder from the project sidebar
- Drag nodes from the palette onto the canvas
- Connect nodes by drawing edges between their ports
- Configure each node's settings
- Save and activate the workflow
Node Types
| Node | Description |
|---|---|
| AI Agent | Run an AI agent with a goal and configuration |
| HTTP Request | Make API calls to external services |
| Condition | Branch based on conditions (if/else) |
| Transform | Transform data between steps |
| Trigger | Start the workflow (webhook, schedule, event) |
| Output | End the workflow and produce a result |
Triggers
| Trigger | Description |
|---|---|
| Webhook | Start from an external HTTP request |
| Schedule | Run on a cron schedule |
| Event | React to platform events (file change, deployment, etc.) |
| Manual | Start manually from the UI |
Data Flow
Data passes between nodes through edges:
- Each node produces an output
- Downstream nodes can reference upstream outputs
- Use the Transform node to reshape data between steps
Example: Auto-Deploy Pipeline
Trigger (push to main)
→ AI Agent (run tests)
→ Condition (tests passed?)
→ Yes: Deploy (to staging)
→ No: Notify (Slack message)