A real-time microservice to stream AI answers word by word
- Context
- The rise of LLMs opened new ways to make sense of survey results: comment summaries, trends, action plans.
- Problem
- An LLM generates its answer word by word, over several seconds. Showing it at the same pace called for a real-time connection all the way to the browser.



