Vizard Agent for podcasters and interview hosts

Asked for

“Analyze the a16z podcast video and extract the 5 most engaging and high-retention 9:16 Shorts. Clip Selection Criteria: - Clip duration should be strictly between 20 and 55 seconds. - Each with a strong hook. - Automatically insert a dynamic visual and audio transition. - Use the "Split Screen" template (Speaker 1 on top, Speaker 2 on bottom). Ensure Auto-Framing keeps faces centered. Apply dynamic "Smart Zoom" (punch-in) on a speaker's face during high-energy moments, punchlines, or loud reactions to emphasize emotion. - Eliminate all dead air and silent pauses. Automatically insert relevant sound effects on key impactful words, jokes, or moments. - Apply dynamic, bold auto-captions but avoid covering the speaker's face. - Each clip ends with the same a16z outro.”

Asked for

“You are editing a 6–7 minute vertical (9:16) interview video into multiple short Instagram Reels, in ONE run. Do not stop to ask me for approval. Plan the clips yourself, then render all of them. Use ffmpeg (and Python/Pillow for text/emoji rendering). ## Context - Location: rooftop of Google's San Francisco office, Bay Bridge in the background, "Google San Francisco" sign on the left. The background is a key selling point, so show it clearly at the start of each clip. - Two people seated: LEFT = Will (the interviewee), RIGHT = me, Yoyo (the interviewer). - Will's face is at roughly x ≈ 36%, y ≈ 49% of the frame. My face is at roughly x ≈ 61%, y ≈ 52%. Verify by grabbing a frame and adjust if needed. - The interview is in English. Audience: Hong Kong, Traditional Chinese speakers, ages 20–40, interested in AI/tech careers. - I've attached an .srt transcript (if provided, use it; otherwis”