Video Generation · Google Flow / Veo 3 · Text / General LLM
88Quality
93%Useful
79Reliability
The prompt
Create an original generative-video prompt for this job: turn one podcast episode into a high-retention social teaser built around a single surprising insight. Inputs: {{episode_topic}}, {{speaker}}, {{key_quote_or_idea}}, {{audience}}, {{platform}}. Make the result commercially usable, visually coherent, and easy for a video model to follow. Specify subject, setting, action, shot order, lens or framing, camera movement, lighting, pacing, transitions, sound or voice-over, and any on-screen text. Preserve logos, products, faces, buildings, interfaces, and reference-image geometry when supplied; do not add unsupported features. Produce a 20-second teaser with cold open, speaker framing, animated captions, b-roll prompts, punch-in timing, music bed, and cliffhanger close. Avoid warped hands, duplicated objects, unreadable text, jumpy continuity, random camera motion, overdone slow motion, and generic AI glow. Keep the CTA legible and reserve safe space for platform UI. Use realistic physics and consistent wardrobe, props, and lighting across shots.