workflow/Workflow Concepts

5.1 Workflow Overview

Pop’s visual Workflow system is a core engine for automating multi-step tasks, organizing node-based processing logic, and building applications.
It allows users to create workflows ranging from simple to extremely complex—without writing code—using drag-and-drop nodes, configurable parameters, and visual connections.

Workflows are ideal for:

  • AI document processing
  • Data transformation
  • Chart generation
  • Knowledge base maintenance
  • Project automation
  • Enterprise internal processes
  • Multi-model collaboration & reasoning
  • Any scenario that requires multi-step processing

This chapter helps you understand what Pop Workflows are, why they are needed, and what they consist of.

🧩 1. What Is a Workflow?

A workflow is a visual flowchart composed of multiple nodes, where each node performs a specific function. Nodes are connected so that the output of one becomes the input of the next, forming a complete automated process.

Example structure:

Start → Upload Document → Chunking → AI Summary → Generate Report → End

Pop automatically executes each step and transfers data between nodes.

🧱 2. Core Components of a Workflow

A complete Pop Workflow consists of:

Workflow
 ├── Nodes
 ├── Edges
 ├── Inputs
 ├── Outputs
 ├── Variables
 ├── Runtime Engine
 └── Layout (if published as an app)

Details below.

🔶 1. Nodes

Nodes are the smallest execution units of a workflow.

Each node performs a function such as:

  • Calling an AI model
  • Generating a PDF
  • Reading an Excel file
  • Making an HTTP request
  • Writing to a database
  • Conditional branching
  • Looping over data
  • Formatting text or JSON
  • Invoking local or remote MCP tools

Nodes contain three types of properties:

  • Inputs
  • Outputs
  • Config

🔁 2. Edges

Edges connect two nodes:

Node A (output) → Node B (input)

They define execution order and data flow.

Edges support:

  • Data flow — passing data forward
  • Control flow — directing process branches (e.g., true/false of a condition)

🧮 3. Workflow Inputs

Workflow inputs are parameters provided at runtime, such as:

  • User-submitted content
  • Runtime variables
  • External parameters passed by an app
  • Form inputs

These values can be referenced by nodes throughout the workflow.

📤 4. Workflow Outputs

After execution, a workflow can output:

  • Model responses
  • File URLs
  • Data objects
  • Chart JSON
  • Final AI reasoning result

These outputs can be:

  • Displayed in apps
  • Rendered in layout manager components
  • Saved to files
  • Used as inputs to subsequent workflows

🔧 5. Workflow Variables

Variables store data generated during workflow execution:

  • System variables
  • Cached node outputs
  • User-defined variables
  • Intermediate values produced by PSL scripts

Variables are accessible across the entire workflow.

🏃 6. Runtime Engine

Pop’s powerful runtime engine is responsible for:

  • Sequential and parallel execution
  • Managing data flow
  • Error capturing
  • Retry & compensation logic
  • Execution context management
  • Multi-model concurrent inference

This engine ensures workflows run reliably.

🖼 7. Workflow Layout (Publishing as an App)

When a workflow is published as an application, it binds to a front-end layout.

Layouts define:

  • Page structure
  • UI components (charts, text, tables, input fields, buttons, etc.)
  • Data binding (workflow output → UI display)

With this, users can build interactive applications without writing front-end code.

🎯 3. Why Use Workflows?

✔ Visualize complex logic

Reduces development cost and improves clarity.

✔ Automate multi-step processes

Boosts efficiency and reduces repetitive tasks.

✔ Enable model collaboration

Combine multiple AI models to achieve advanced reasoning.

✔ Implement business workflows

Examples: auto-report generation, automated customer service, data pipelines.

✔ Modular & reusable

Nodes and sub-workflows can be reused across projects.

🧠 4. Relationship Between Workflows and Programming

Workflow ≠ Code
Workflow = Visual representation of logic

For advanced logic, Pop provides PSL (Pop Script Language) which supports:

  • Array processing
  • String operations
  • Data transformation
  • Conditional structures
  • Complex control flows

Workflows + PSL = Complete automation capability.

📌 5. Summary

Pop Workflow is a comprehensive visual automation engine containing:

Feature Description
Node System Performs different tasks
Edge System Controls flow & data passing
I/O System Defines workflow inputs/outputs
Variable System Manages runtime data
Layout System Enables publishing as apps
PSL Scripting Supports complex logic
Previous page
Knowledge Base Management & Index Maintenance
Next Page
Workflow Editor Overview