QuestSharp

A goal-oriented conversation flow framework built on top of Microsoft's Semantic Kernel Process Framework, enabling structured and intelligent human-LLM interactions.

Overview

QuestSharp is a framework that combines the power of Microsoft's Semantic Kernel Process Framework with goal-oriented conversation flows. It allows developers to create sophisticated AI assistants that can:

• Handle complex multi-step conversations with natural language
• Maintain context and state across interactions
• Validate user inputs with custom validation rules
• Route conversations between different goals based on user intent
• Process structured data collection through conversation

Key Features

• Goal-Based Architecture: Define conversation flows as a series of goals with required fields and validation rules
• Natural Conversation Flow: LLM-powered natural language processing for fluid interactions
• Field Validation: Built-in validation system with custom validators for each field
• Flexible Routing: Connect goals together with intent-based transitions and context preservation
• Process Integration: Seamless integration with Semantic Kernel's Process Framework
• State Management: Maintain conversation history and context across goal transitions
• Error Recovery: Graceful handling of validation failures and conversation errors

Architecture

QuestSharp is built on several key components:

1. Goals: Core building blocks representing conversation objectives

   • Name and description for clear identification
   • Required fields with validation rules
   • Opening messages for conversation initiation
   • Connections to other goals for transitions

2. Process Steps:

   • ConsoleStep: Handles user input/output
   • GoalPromptStep: Processes conversation with LLM
   • ValidationStep: Validates gathered information
   • GoalTransitionStep: Manages transitions between goals
   • RenderResponseStep: Formats and delivers responses

3. Services:

   • QuestService: Orchestrates the conversation flow and maintains state

Getting Started

1. Define your goals using the builder pattern:

var orderGoal = new GoalBuilder()
    .WithName("OrderProcess")
    .WithDescription("Process customer orders")
    .WithOpener("Welcome! Let me help you with your order.")
    .AddField("quantity", "Number of items", val => int.Parse(val.ToString()) > 0)
    .Build();

2. Connect goals to create conversation flows:

orderGoal.Connect(cancelGoal, "cancel", handOver: true);
orderGoal.Connect(menuGoal, "menu", handOver: true);

3. Initialize the service and start processing:

var chain = new QuestService(kernel, initialGoal);
await chain.ProcessInputAsync(userInput);

Requirements

• .NET 8.0 SDK or higher
• Azure OpenAI or OpenAI API access

Documentation

For detailed documentation on the Semantic Kernel Process Framework that powers QuestSharp, visit: Process Framework Documentation

Acknowledgments

This project is inspired by GoalChain, reimagined and rebuilt using Microsoft's Semantic Kernel Process Framework.

License

MIT License - See LICENSE file for details