Ollama chatClient streaming response

[khi195]

Hey, been looking into this for a few days.

When I stream back a response, I get words broken into chunks making it difficult to render on the frontend. That response chunking from spring AI chat client is sending back "cup of gran ulated sugar" instead of "cup of granulated sugar"

I'm using Kotlin, and this is the code

fun doSomething(@RequestBody request: Request): Flux<AssistantMessage> {
        val stream = ChatClient.prompt(request.message).stream()


        return stream.chatResponse().map { it.result.output }
    }

What am I missing ?

Many thanks in advance.

[khi195]

replace the above kotlin with the following, and it seems to work.

@PostMapping(produces = [MediaType.TEXT_EVENT_STREAM_VALUE])
    fun doSomething(@RequestBody request: Request): Flux<ChatResponse> {
        val stream = ChatClient.prompt(request.message).stream()


        return stream.chatResponse()
    }