Understanding Workflows (Flows)

Flows let you chain agents together into automated pipelines. Instead of manually running each step, you define the sequence once and let it run on its own.

What is a Flow?

A flow is a visual diagram that defines:

1. When to start (trigger)
2. What to do (agents and actions)
3. How to decide (conditions)
4. Where to go next (connections)

┌──────────┐      ┌─────────────┐      ┌─────────────┐
│  Trigger │─────▶│ Agent: Find │─────▶│ Agent: Write│
│  (Daily) │      │   Papers    │      │   Summary   │
└──────────┘      └─────────────┘      └─────────────┘

Common Use Cases

Flow Type	Example
Research Pipeline	Find papers → Summarize → Create ticket
Code Review	Analyze code → Check security → Wait for approval → Merge
Data Processing	Fetch data → Clean → Analyze → Generate report
Ticket Automation	New ticket → Triage → Assign → Notify team

Building Blocks

Triggers

Every flow starts with a trigger:

Trigger	When it fires
Manual	You click "Run"
Schedule	At specific times (daily, hourly, etc.)
Ticket Created	When a new ticket is created
Ticket Updated	When a ticket changes

Nodes

Nodes are the steps in your flow:

Node Type	What it does
Agent	Runs an AI agent with input/output
Condition	Routes based on a test (if/else)
Wait	Pauses for human approval or input
Parallel	Splits into multiple branches
Merge	Joins branches back together
Ticket Action	Creates or updates tickets
Transform	Modifies data between steps

Connections

Lines between nodes show the execution order. For condition nodes, you'll have two paths: one for "yes" and one for "no".

Creating Your First Flow

1. Go to Flows in the sidebar
2. Click New Flow
3. Drag a Trigger node onto the canvas
4. Drag an Agent node and connect it
5. Configure each node's settings
6. Click Save
7. Click Run to test

Passing Data Between Nodes

Each node can access data from previous nodes using template syntax:

{{trigger.input}}           // What started the flow
{{previousNode.output}}     // Output from the last node
{{nodes.myAgent.output}}    // Output from a specific node

Controlling Execution

While a flow runs, you can:

• Pause it to inspect progress
• Resume after pausing
• Cancel to stop entirely
• Inject data when a Wait node needs input

Best Practices

Start simple. Build a linear flow first, then add conditions and parallelism.

Name nodes clearly. "Summarize Papers" is better than "Agent 1".

Test incrementally. Run after adding each node to catch problems early.

Handle errors. Add condition nodes to check for failures and respond appropriately.

Avoid infinite loops

If you create cycles in your flow, make sure there's a condition that eventually exits the loop.

Next Steps

• Research Pipeline - Build a multi-agent flow
• Flow Editor reference
• Node types in detail
• Example flows