ranade-oss fixes the handoff from live chat to a finished answer
A small hotfix makes streamed assistant replies land as completed responses once their saved output is ready.
The fork now explicitly handles the last step of a live assistant exchange: turning saved streamed text into the finished answer a reader sees.
That matters because a chat experience can feel unreliable if a response remains stuck looking in-progress after the words have already arrived. This is a narrow change, but it addresses a very visible moment in any legal-AI workflow: whether the assistant actually closes the loop.
So what Teams assessing this fork's chat experience should care because dependable response completion is basic to trusting an assistant in day-to-day legal work.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?