AgentBuilder provides four types of applications:

• Chatbot
• Text Generator
• Agent
• Workflow

---

Differences Between Text Generator and Chatbot

Feature	Text Generator	Chatbot
WebApp Interface	Form + Results	Chat-based
WebAPI Endpoint	completion-messages	chat-messages
Interaction Mode	One question, one answer	Multi-turn conversation
Streaming Results	Supported	Supported
Context Preservation	Per session	Continuous
User Input Form	Supported	Supported
Datasets and Plugins	Supported	Supported
AI Opening Remarks	Not supported	Supported
Example Scenarios	Translation, judgment, indexing	Chatting

---

Application Types Overview

• Chatbot

  • Built on LLM for conversational interaction.
  • Supports two orchestrate methods:
    • Basic orchestrate: Simple configuration, no need to modify built-in prompts.
      • 
      • Edit page:
    • Chatflow orchestrate: Arranges chatbots as workflows, highly customizable, allows editing built-in prompts.
      • 
      • Edit page:

• Text Generator

  • Automatically generates high-quality text based on user prompts (e.g., summaries, translations).
    • 
    • Edit page:

• Agent

  • Uses LLMs to autonomously plan goals, break down tasks, invoke tools, iterate, and complete tasks without human intervention.
    • 
    • Edit page:

• Workflow

  • Orchestrates multiple tasks and models into a workflow for complex scenarios.
    • 
    • Edit page:

---