Currently BoltAI renders markdown and code block etc. for a user message (prompt) but I think it doesn't make sense. For 2 reasons:
  1. It's awkward to render markdown and code block on a blue background (for user message), and it ends up bringing many readability issue (low contrast etc.)
  2. The user prompt is the input to the LLM and I believe we should render it as plaintext. So that if we find any issue or mistake in the prompt, we can spot it right away. Rendering in a different format removes this ability. Worse, it might even renders it incorrectly
Another solution is to ditch the "iMessage style" and make it similar to ChatGPT web app where the user and AI messages are on the same side. This has another UX issue is that it's harder to distinguish between user & AI messages.
Let me know what you think in the comment ↓