stream
open fun streamInference(messages: List<Message>, tools: List<Tool>): <Error class: unknown class><LlmInferenceStreamEvent>
Stream one inference without delegating tool execution to the provider.
Provider adapters should override this method to emit an assembled Complete event containing any requested tool calls. The default bridge keeps existing LlmMessageStreamer implementations usable: their raw stream is replayed as Content and summarized as a terminal assistant message. Such legacy adapters retain their provider-managed tool behavior and cannot expose assembled tool calls to DefaultStreamingToolLoop until they override this method.