Real-time Stream-ing of Responses to SfAIAssistView
Hello Syncfusion Team and Community,
I am building a Blazor application that integrates with a local Ollama server to create an AI assistant using the SfAIAssistView component. The component is excellent for handling user prompts and displaying a finished response.
I am trying to implement a stream-ing response feature where the AI's reply is rendered in real time, character by character or chunk by chunk, as it is generated by the server. This provides a much better user experience, as it feels more dynamic and responsive.
I would like to request the addition of a feature that allows developers to:
Stream responses directly to
SfAIAssistView.Provide a method, such as
args.AppendResponseAsync(string chunk)orargs.SetResponseStream(IAsyncEnumerable<string> stream), within theAssistViewPromptRequestedEventArgsthat enables real-time rendering.
My current code uses the PromptRequested event to call my IChatInferenceService:
@page "/chat"
@attribute [Authorize]
@using Syncfusion.Blazor.InteractiveChat
@using ChatMessage = Microsoft.Extensions.AI.ChatMessage
@inject IChatInferenceService OllamaAIService
<SfAIAssistView PromptRequested="@PromptRequest"></SfAIAssistView>
@code {
private async Task PromptRequest(AssistViewPromptRequestedEventArgs args)
{
args.Response = await OllamaAIService.GenerateResponseAsync(new ChatParameters
{
Messages = [ new ChatMessage(Microsoft.Extensions.AI.ChatRole.User, args.Prompt) ]
});
}
}
