“sse”

Streaming Inference with FastAPI and Server-Sent Events

Turn a blocking LLM call into an incremental, observable stream of tokens with FastAPI StreamingResponse and Server-Sent Events — including cancellation and backpressure.