Build Application, AI Agent, Workflows

Build powerful AI applications with an intuitive drag-and-drop interface — no coding required. ClickAI Workflow Editor

Table of Contents

· [Introduction](#introduction)

· [Application Types](#application-types)

· [Workflow & Chatflow Editor](#workflow--chatflow-editor)

· [App Toolkit — Extended Features](#app-toolkit--extended-features)

· [Build Process](#build-process)

· [Best Practices](#best-practices)

Introduction

ClickAI Studio provides a visual development environment for building AI applications — from simple chatbots to complex automation workflows. The platform supports multiple application types, each optimized for different use cases.

💡 TIP: You don't need programming experience to start building AI applications on ClickAI. The drag-and-drop interface enables anyone to create professional AI applications.

Application Types

1. Chatbot & Agent

AI Agent at Click AI built by Agentic Architecture, it capable of autonomous reasoning, planning, and using tools to complete complex tasks.Interactive conversational application using Large Language Models (LLM) for Q&A.

Feature	Description
Communication	Multi-turn conversation
Input	Text messages from users
Output	AI-generated text responses
Use Cases	Customer support, virtual assistant, FAQ bot

Feature	Description
Capability	Auto-select and use tools
Reasoning	Chain-of-thought reasoning
Tools	Web search, Code interpreter, API calls, ...
Use Cases	Market research, data analysis, automation

⚠️ IMPORTANT: Agents can call multiple tools in a single processing turn. Ensure correct permissions are configured for connected tools.

2. Workflow

Automated processing pipeline — sequential execution from start to finish, no conversational interaction.

Feature	Description
Execution	Sequential, batch processing
Input	Input variables
Output	Final processing result
Use Cases	ETL data, report generation, batch processing

3. Chatflow

Combines the power of Workflow with a conversational interface — enables building complex chatbots with multi-layer processing logic.

Feature	Description
Communication	Conversation + branching logic
Power	All Workflow nodes + chat context
Use Cases	Complex customer service, step-by-step guidance

Workflow & Chatflow Editor

Core Nodes

ClickAI provides a rich node system for building processing logic:

🔄 Flow Control

Node	Function
Start	Entry point, define inputs
End	Exit point, return results
IF/ELSE	Conditional branching logic
Iteration	Loop through item lists
Loop	Loop with stop condition
Parameter Extractor	Extract parameters from natural language

🧠 Model

Node	Function
LLM	Call Large Language Models (GPT, Claude, Gemini, ...)
Knowledge Retrieval	Query Knowledge Base
Question Classifier	Classify user questions

🔧 Tools

Node	Function
Code	Run custom Python / JavaScript code
HTTP Request	Call external APIs
Template	Format output with Jinja2 templates
Variable Assigner	Assign values to variables

📤 Output

Node	Function
Answer	Reply directly to user (Chatflow)
End	Return final result (Workflow)

Variables & Data

ClickAI supports the following variable types:

· Input Variables: User-provided input variables

· Environment Variables: Environment variables (API keys, secrets)

· Conversation Variables: Session-scoped variables

· System Variables: sys.query, sys.user_id, sys.conversation_id, ...

App Toolkit — Extended Features

Optional features to enhance your applications:

💬 Conversation Opener

Set up welcome messages and initial question suggestions for users.

🔄 Follow-up

Automatically suggest follow-up questions based on conversation context.

🔊 Text to Speech (TTS)

Convert text responses to natural speech.

🎤 Speech to Text (STT)

Allow users to input via voice.

📎 File Upload

Allow file uploads as input:

File Type	Default Limit
Image	10 MB
Document	15 MB
Audio	50 MB
Video	100 MB

📝 Citations & Attributions

Display source citations from Knowledge Base, helping users verify information.

🛡️ Content Moderation

Filter inappropriate content with 3 methods:

7. OpenAI Moderation: Dedicated moderation model

8. Keywords: Blocked keyword list

9. API Extension: Custom content filtering API

✏️ Annotation Reply

Create a library of standard responses, bypassing AI when encountering annotated questions.

Build Process

Step	Action	Details
1	Define Use Case	Clearly identify purpose and target audience
2	Choose App Type	Chatbot, Agent, Workflow, or Chatflow
3	Configure Model	Select LLM, write System Prompt
4	Connect Knowledge	Upload documents to Knowledge Base
5	Add Tools	Connect APIs, plugins, custom code
6	Debug & Preview	Test directly in Studio
7	Publish	Deploy via Web App, API, or Embed

Best Practices

💡 TIP: **Effective System Prompts:** Define clear role and response style, set explicit scope boundaries, use few-shot examples.

🚨 WARNING: **Important Notes:** Always test thoroughly before publishing to production. Set up Content Moderation for public-facing apps. Monitor token usage to control costs.

📖 Next: [Publish — Deploy Applications]