Streaming endpoint currently requires all clients to implement the logic of collecting all chunks to assemble the tool call. It would be nice to have a helper function on the LLM app that does this on behalf of the client so it doesn't have to be implemented multiple times.