ai-agents-chain-streaming-llmchain-s-stream-method-is-not-implemented-causin-24077076

THEN Override the stream method in a custom chain that extends LLMChain. In the stream method, call self.prep_prompts with the input, then yield from self.llm.stream with the prepared prompt. This allows chunk-by-chunk streaming output.

class MyChain(LLMChain):
    def stream(self, input, config=None, run_manager=None, **kwargs):
        prompts, stop = self.prep_prompts([input], run_manager=run_manager)
        yield from self.llm.stream(input=prompts[0], config=config, **kwargs)