Basic Chatbot

This example demonstrates how to create a basic chatbot workflow that takes user input, processes it through a language model, and returns a response.

Workflow Structure

Here's a basic chatbot flow running in action:

Step-by-Step Instructions

Step 1: Add QueryNode

1. Add a QueryNode to the canvas
2. Configure it with a default query: "Hello, how can you help me?"
3. This node will serve as the entry point for user input

Step 2: Add LanguageModelNode

1. Add a LanguageModelNode to the canvas
2. Configure it:
   • Service: OpenAI (or your preferred provider)
   • Model: gpt-3.5-turbo or gpt-4
   • System Prompt: "You are a helpful AI assistant."
   • Temperature: 0.7
   • Max Tokens: 500

Step 3: Add ResponseNode

1. Add a ResponseNode to the canvas
2. This node will output the final response to the user

Step 4: Connect the Nodes

1. Connect QueryNode's "query" output → LanguageModelNode's "query" input
2. Connect LanguageModelNode's "response" output → ResponseNode's "response" input

Configuration Details

QueryNode Configuration:
  - query: "Hello, how can you help me?"

LanguageModelNode Configuration:
  - service: "openai"
  - model: "gpt-3.5-turbo"
  - system_prompt: "You are a helpful AI assistant."
  - temperature: 0.7
  - max_tokens: 500

ResponseNode Configuration:
  - (No configuration needed)

Testing the Workflow

You can also start with pre-built templates from the templates page:

1. Click the "Execute" button
2. Watch as the workflow processes:
   • QueryNode outputs the user query
   • LanguageModelNode processes the query and generates a response
   • ResponseNode displays the final response
3. Check the ResponseNode to see the AI's response

Enhancing the Chatbot

You can enhance this basic chatbot by adding:

• Intent Classification: Add an IntentClassificationNode to categorize user queries
• Knowledge Base: Add a KnowledgeBaseRetrievalNode to provide context-aware responses
• Web Search: Add a WebSearchNode to fetch real-time information
• Conditional Logic: Add a ConditionalNode to handle different conversation paths