Short answer: yes—render the agent’s text as Markdown in your UI.
In Chat Mode, agent replies are delivered as text via the agent_response event/callback; formatting is up to your client. (elevenlabs.io)
To display headings, lists, code blocks, etc., use the ElevenLabs UI “Response” component, which is a streaming Markdown renderer (built on Streamdown) and “supports full markdown syntax including code blocks, lists, tables, and more.” (ui.elevenlabs.io)
You can also steer the LLM to produce well-structured Markdown by specifying formatting expectations in the agent’s system prompt; the Prompting Guide recommends using Markdown headings and bulleted lists in instructions. (elevenlabs.io)
Practical recipe:
Enable Chat Mode and handle agent_response to capture the text message. (elevenlabs.io)
Render that text with the UI Response component so Markdown is parsed and displayed as headings and lists. (ui.elevenlabs.io)
tawny scarab
cloud lance