Introduction

Some prompts don't try to look "AI." They try to look like a forgotten MiniDV tape your friend shot on a sticky Seoul evening — and that is exactly why this one hits.

Credit to

for the original drop. Full prompt preserved below in a Code Box so it still lives here if the post ever vanishes.

Why it works

It locks one consistent subject, a lived-in residential lane (not tourist Seoul), and a ruthless early-2000s DV camera language: handheld shake, autofocus hunting, faded color, accidental zooms. Then it storyboards six five-second beats — fruit shop bit, dog hello, leaves in hair, "Again?" — so the model has a timeline instead of vibes alone.

Natural location audio only. No music. No logos. Continuity rules that ban the usual gen-AI tells. Warm, casual, nostalgic.

Original source

The prompt (plaintext archive)

Create a 30-second, 1080p, 16:9 ultra-realistic early-2000s DV home video of an ordinary summer evening in Seoul. No reference image.

Subject: Same young Korean woman throughout—early 20s, natural realistic skin, minimal makeup, long dark-brown hair in a low messy bun with loose strands. Soft-yellow sleeveless knit top, loose light-gray trousers, white retro sneakers, dark-green shoulder bag, silver bracelet. Keep face, hair, outfit, proportions, and accessories perfectly consistent.

Setting: Quiet older Seoul residential lane with small apartments, potted plants, bicycles, old walls, utility poles, wires, a tiny snack shop, drinking fountain, and leafy trees. Lived-in, peaceful, authentic; no brands, ads, tourist locations, or polished commercial activity.

Camera: Raw friend-recorded early-2000s consumer DV footage—strong handheld shake, imperfect framing, autofocus hunting, exposure shifts, motion blur, faded colors, soft digital detail, mild noise, accidental zooms. No stabilization, gimbal, drone, or cinematic shots.

Timeline:

- 00:00–00:05: Leaves home with reusable shopping bag, locks door, checks keys, fixes bun, smiles at camera, walks down lane.
- 00:05–00:10: Buys cold fruit in a paper cup from a tiny snack shop. Takes a bite, reacts to sour taste, makes a funny face, laughs, offers fruit toward camera.
- 00:10–00:15: Walks with fruit. A small dog notices her, wags, and she crouches to let it sniff her hand. Dog follows briefly, then returns.
- 00:15–00:20: Dry leaves fall from a tree onto her shoulder and hair. She looks up, brushes them off, awkward camera zooms toward her face. She says, "Again?"
- 00:20–00:25: Sits briefly on a low wall, finishes fruit while a bicycle passes and a neighbor waters plants. Breeze moves her hair and clothes; she wipes her hands and adjusts her bag.
- 00:25–00:30: Walks toward the end of the lane, turns back with a playful smile and says, "Let's go." Walks around the corner; camera follows; abrupt cut to black.

Audio: Natural location sound only—birds, insects, footsteps, bicycle bells, distant scooters, leaves, shop sounds, dog movement, neighborhood voices, and camera-handling noise. No music, narration, subtitles, captions, logos, watermarks, or artificial effects.

Continuity: Realistic dog and environmental movement. No duplicated people, distorted faces, broken hands, extra fingers, teleportation, or outfit changes. Warm, casual, feminine, playful, intimate, nostalgic forgotten-summer-memory feeling.

Pro Tip

Keep the Subject paragraph untouched across regenerations if you want identity continuity. If your video model struggles with dialogue, keep the spoken lines short ("Again?" / "Let's go.") and lean on the action beats.

Conclusion

Paste, generate, and chase that "this doesn't look like AI" feeling — friend-cam energy, one summer lane, thirty honest seconds.

—Aurelia ♡