---
title: "Discussion on \"Streaming LLM Tokens Over SSE in Production\" | Hashnode"
description: "Discussion on \"Streaming LLM Tokens Over SSE in Production\". Why ChatGPT and Claude's \"typing\" effect is just SSE: the wire format, why EventSource can't call these APIs directly, and the backend relay pattern with a working Spring Boot example"
url: "https://hashnode.com/posts/streaming-llm-tokens-over-sse/6a708ddccf7091ad66b332ed"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Streaming LLM Tokens Over SSE in Production](https://buildwithmostafa.hashnode.dev/streaming-llm-tokens-over-sse?utm_source=hashnode&utm_medium=feed)

Why ChatGPT and Claude's "typing" effect is just SSE: the wire format, why EventSource can't call these APIs directly, and the backend relay pattern with a working Spring Boot example

Aug 3 · 18 min read · buildwithmostafa.hashnode.dev  
Read the article

-   [#llm](https://hashnode.com/tag/llm)
-   [#server-sent-events](https://hashnode.com/tag/server-sent-events)
-   [#sse](https://hashnode.com/tag/sse)
-   [#springboot](https://hashnode.com/tag/springboot)
-   [#generative-ai](https://hashnode.com/tag/generative-ai)
-   [#api](https://hashnode.com/tag/api)
-   [#ai](https://hashnode.com/tag/ai)

## Written by

[Mostafa Mahmoud](https://hashnode.com/@buildwithmostafa)

@buildwithmostafa · Cairo, Egypt

  
[View profile](https://hashnode.com/@buildwithmostafa)

Software Engineer | Building Robust Backends & Intelligent Workflows

Software Engineer specializing in Java, Spring Boot, and Camunda 8. I write about how modern backend systems are designed, integrated, and orchestrated in production.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
