{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/91bb5bae871749c784251133ee14becc\" frameborder=\"0\" width=\"1280\" height=\"960\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":960,"width":1280,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":960,"thumbnail_width":1280,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/91bb5bae871749c784251133ee14becc-82741b41bfa9de1f.gif","duration":282.918,"title":"Exploring the Zero Touch AI Avatar Pipeline 🚀","description":"Hi team, in this video, I walk you through the Zero Touch AI Avatar Pipeline I built, showcasing how we process a YouTube URL and generate an avatar with a high-resolution thumbnail and cloned audio using Grok and fal.ai. I specifically engineered an asynchronous loop to handle API timeout limits effectively, ensuring smooth GPU rendering. After processing, we log the data into a SuperBase SQL database and I demonstrate the frontend I created to visualize the output. I will also share the video through Dropbox and my GitHub profile for further reference. Please take a look at the demo and let me know your thoughts!"}