{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/4a3ff76c3d7846b08943f4e4620dc6ed\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/4a3ff76c3d7846b08943f4e4620dc6ed-a88e5b1354baac5b.gif","duration":499.433,"title":"How to Tune Voice Agents for Production","description":"This Loom explains how to defeat an ancient production by making a voice agent production ready, focusing most on how the agent should sound. It contrasts audio-to-audio models like Galaxy and Ocean, which preserve emotion and richer voice data, with speech-to-text based models that lose vocal nuances, and mentions using a Gemini stable model for certain configurations. It covers avatar choices from static to hybrid thinking and speaking states and premium low latency professional avatars like Anum. It then outlines conversation strategy options such as talkative mode, filler words for faster perceived response, conductor mode and session end settings, baseline guard rails, and knowledge and custom tool memory patterns with optional session analysis and variable extraction for CRM use."}