Street Interview With a Hook Question (Veo 3)
The viral man-on-the-street format for Veo 3: physically anchored camera, one question, one answer, and native audio doing the production work.
Prompt variables
7 variables detected — fill them and Copy inserts your values (0/7 filled).
Street interview video, {{city_environment}}, {{time_of_day}}. A {{interviewer_description}} holds a handheld microphone with a small cube logo flag — the camera is a second person standing two steps away at chest height, slightly handheld with natural micro-shake. The interviewer stops {{interviewee_description}} walking past and asks: "{{hook_question}}" A half-second beat — the interviewee's expression shifts to {{reaction}} — then they answer: "{{surprising_answer}}"
Framing: loose two-shot, both faces visible, busy sidewalk life continuing behind them with pedestrians crossing the background naturally out of focus. Audio: city ambience (traffic hum, distant chatter), both voices clear and close thanks to the visible microphone, no music. Style: authentic social-video realism, natural color, no cinematic grade.
Format notes:
- Anchor the camera as a PHYSICAL thing ("a second person standing two steps away") — Veo 3 handles "someone is filming this" far better than abstract terms like "POV" or "handheld footage".
- One question + one short answer fits the ~8-second clip. Write both lines to be speakable in one breath each; longer dialogue slides out of sync. For a full episode, generate one exchange per clip and jump-cut like real street-interview channels do.
- The visible microphone is the audio alibi: it justifies clean voice-over-ambience mixing and instantly codes the genre.
- The reaction beat between question and answer is where the entertainment lives — specify it ("bursts out laughing", "goes dead serious", "looks around to check who's listening").
- "No music" keeps the clip editable; Veo otherwise loves to bake in a score you can't remove.
Proof of work
Example outputs
Prompt as run: Street interview video, busy Copenhagen shopping street, overcast afternoon. A woman in a yellow raincoat holds a handheld microphone with a small cube logo flag — the camera is a second person standing two steps away at chest height, slightly handheld with natural micro-shake. She stops a man in a gray wool coat walking past and asks: "What's one thing you own that you'd never sell?" A half-second beat — his expression shifts to a slow grin — then he answers: "My grandfather's broken accordion. It doesn't even play." Result notes: dialogue synced cleanly in 3 of 4 generations; ambience mixed under the voices exactly as specified. The micro-shake phrasing produced believable second-person framing — earlier attempts with "POV handheld" had given floating drone-like moves.
Comments (0)