BasicChat-10ConversationHistory

BasicChat-10ConversationHistory

Understanding Chat Message Lists and GetResponseAsync

This example demonstrates the correct way to handle conversation history when using GetResponseAsync() with Microsoft.Extensions.AI.

Common Issue: Adding Response to Conversation

When building a chatbot with conversation history, developers often encounter this error:

List<ChatMessage> conversation = new() { ... };
var response = await client.GetResponseAsync(conversation);
conversation.Add(response.Messages);  // ❌ ERROR: response doesn't have .Messages property

Why This Error Occurs

The GetResponseAsync() method returns a response object that has:

• ✅ .Text property - Contains the text response from the AI
• ✅ .Message property - Contains a single ChatMessage object
• ❌ NO .Messages property (plural)

The Correct Pattern

Here's how to properly maintain conversation history:

// 1. Initialize conversation with system message
List<ChatMessage> conversation = new()
{
    new ChatMessage(ChatRole.System, "You are a helpful assistant")
};

// 2. Add user message
conversation.Add(new ChatMessage(ChatRole.User, userQuestion));

// 3. Get AI response
var response = await client.GetResponseAsync(conversation);

// 4. Add assistant response to history - THREE CORRECT WAYS:

// Option A: Using response.Text (most common)
conversation.Add(new ChatMessage(ChatRole.Assistant, response.Text));

// Option B: Using response.Message (contains the full message object)
conversation.Add(response.Message);

// Option C: Using the Message property directly for more control
var assistantMessage = new ChatMessage(ChatRole.Assistant, response.Text);
conversation.Add(assistantMessage);

Key Properties of Response Object

When you call GetResponseAsync(), the response object provides:

Property	Type	Description
.Text	string	The text content of the AI's response
.Message	ChatMessage	A ChatMessage object with Role=Assistant and the response text
.Contents	IList<AIContent>	Collection of content items (text, images, etc.)
.FinishReason	ChatFinishReason?	Reason why the response ended

Running This Example

1. Ensure Ollama is running:

   ollama serve

2. Pull the required model:

   ollama pull phi4:latest

3. Run the application:

   dotnet run app.cs

4. Type "conversation" to start chatting, then enter your questions.

Alternative: Using GetStreamingResponseAsync

For streaming responses (showing text as it arrives), use a different pattern:

var sb = new StringBuilder();
await foreach (var item in client.GetStreamingResponseAsync(conversation))
{
    sb.Append(item);
    Console.Write(item.Contents[0].ToString());
}
conversation.Add(new ChatMessage(ChatRole.Assistant, sb.ToString()));

See BasicChat-05AIFoundryModels for a complete streaming example.

Common Mistakes to Avoid

❌ Wrong: conversation.Add(response.Messages) - .Messages doesn't exist
❌ Wrong: conversation.Add(response) - Can't add response object directly
❌ Wrong: conversation.Add(response.Text) - Can't add string directly

✅ Correct: conversation.Add(new ChatMessage(ChatRole.Assistant, response.Text))
✅ Correct: conversation.Add(response.Message)

See Also

• Microsoft.Extensions.AI Documentation
• BasicChat-01MEAI - Simple single-turn conversation
• BasicChat-05AIFoundryModels - Streaming conversation with history