06.3 nodes

Workflow Nodes

Workflow nodes are the building blocks of every BindAI workflow.

Each node performs one specific operation before transferring execution to the next node.

By combining different node types, workflows can model anything from simple AI conversations to complex enterprise automation.

---

What is a Node?

A node represents a single execution step.

Conceptually:

Input

↓

Node

↓

Output

Every workflow is a collection of connected nodes.

---

Workflow Graph

Nodes are connected to form a directed graph.

Start

↓

Agent

↓

Condition

↓

Tool

↓

End

Execution moves from one node to the next according to the workflow structure.

---

Node Lifecycle

Every node follows the same execution lifecycle.

Receive Context

↓

Execute

↓

Update Context

↓

Continue

The workflow context carries information between nodes.

---

Built-in Node Types

BindAI includes several built-in node types.

Node	Purpose
Start	Entry point
End	Workflow completion
Agent	Execute an AI agent
Condition	Branch execution
Loop	Repeat execution
Parallel	Execute multiple branches
Join	Synchronize parallel branches
Human Task	Pause for human approval
Subworkflow	Execute another workflow (future expansion)

Each node has a single responsibility.

---

Start Node

Every workflow begins with a Start node.

Start

↓

Next Node

The Start node initializes workflow execution.

---

End Node

The End node completes workflow execution.

Previous Node

↓

End

Once the End node executes, the workflow finishes.

---

Agent Node

Agent nodes execute AI agents.

Workflow

↓

Agent

↓

Agent Result

The agent can read workflow variables and write new values back into the execution context.

---

Condition Node

Condition nodes create branching logic.

Condition

↙       ↘

True     False

Only one branch continues based on the evaluated predicate.

---

Loop Node

Loop nodes repeat execution until a condition changes.

Loop

↓

Body

↓

Condition

↓

Repeat or Exit

Loops are useful for iterative processing.

---

Parallel Node

Parallel nodes begin multiple execution branches.

Parallel

↙     ↓     ↘

A     B     C

Each branch executes independently.

---

Join Node

Join nodes synchronize parallel execution.

Branch A

↓

Join

↑

Branch B

Execution continues only after all required branches have completed.

---

Human Task Node

Human task nodes pause execution until manual approval is received.

Workflow

↓

Human Task

↓

Waiting

↓

Resume

This enables human-in-the-loop workflows.

---

Subworkflow Node

A subworkflow node delegates execution to another workflow.

Conceptually:

Main Workflow

↓

Subworkflow

↓

Return

↓

Continue

This promotes reuse of common workflow logic.

---

Context Sharing

All nodes operate on the same workflow context.

Node A

↓

Workflow Context

↓

Node B

This allows data to flow naturally through the workflow.

---

Variables Between Nodes

Nodes communicate through workflow variables.

Example:

Agent

↓

customer_name

↓

Condition

↓

Email Agent

Each node may read existing variables and create new ones.

---

Custom Nodes

Developers can create custom nodes by extending the base workflow node.

Conceptually:

class MyNode(
    WorkflowNode,
):
    ...

Custom nodes integrate seamlessly with the execution engine.

---

Node Connections

Each node determines which node executes next.

Node

↓

Next Node

↓

Next Node

Some nodes, such as conditions and parallel nodes, may have multiple outgoing connections.

---

Node Independence

Nodes should remain independent whenever possible.

A well-designed node:

• performs one task
• reads required context
• updates context
• transfers execution

Keeping responsibilities small improves reuse and maintainability.

---

Best Practices

• Give every node a single responsibility.
• Pass information through workflow variables.
• Keep custom nodes focused and reusable.
• Avoid embedding business logic inside workflow orchestration.
• Prefer composing workflows from multiple simple nodes instead of a few complex ones.
• Use descriptive node identifiers for easier debugging.

---

Summary

Workflow nodes are the execution units of BindAI workflows.

Each node performs one clearly defined operation while sharing information through the workflow context. By combining different node types, developers can build reliable AI workflows that are modular, readable, and easy to extend.