GoodCase.aiAI News · Radar↗AI Map↗Course · LearnPrompt↗
ProjectsGoodCase.aiAI News Radar ↗AI Map ↗LearnPrompt ↗
GoodCase.aiGoodCase.aiAI Prompt Case Collection
GoodCase.ai
RankingsCasesSkillsFavoritesSubmitDailyJoin us
中文EN
Agent Access →

Projects

AI News · Radar ↗AI Map ↗Course · LearnPrompt ↗
GoodCase.aiGoodCase.aiAI Prompt Case Collection
RankingsCasesSkillsFavoritesSubmitDailyJoin us
中文EN
Agent Access →

Loading…

GoodCase.ai · AI Prompt Case Collection
DSH PluginsChangelogFeedback
ModelsCreatorsSeedance 2.5Seedance 2.0GPT Image 2GeminiNano Banana ProClaudeClaude Opus 5.5Grok ImagineVeoKlingQwen ImageGPT Image 1.5MiniMax H3

Tokyo Street Food Phone Vlog

Creator · saniaspeaks_

AI VideoX2026-08-15Added 08/19/2026Independently retestedSeedance 2.5
View original source ↗
Source ♥ 291

Prompt

Ultra-realistic smartphone food vlog set in a charming neighborhood in Tokyo, Japan. The entire video must look like it was genuinely recorded on a modern flagship smartphone by a real female creator, not like a cinematic production or AI-generated video. 0–5 seconds: A young Japanese woman walks through a narrow Tokyo side street in the early evening, holding her smartphone at arm’s length and casually talking to the camera. Traditional Japanese shopfronts, vending machines, bicycles, small signs, utility wires, pedestrians and warm storefront lights create an authentic neighborhood atmosphere. Natural handheld movement and slight framing imperfections. 5–10 seconds: She notices a tiny hidden dessert café and turns the camera toward its entrance with an excited reaction. She walks inside while continuing the vlog. The camera naturally adjusts exposure from the bright street to the warmer interior. 10–16 seconds: Show the staff preparing a fresh strawberry cream dessert. Realistic strawberries, whipped cream, soft cake layers and small details of the preparation. The creator moves the phone closer for a natural food close-up while autofocus subtly shifts between the dessert and her hand. 16–22 seconds: She receives the dessert, holds it directly toward the phone camera, then takes a spoonful and tastes it. Her reaction should be spontaneous and believable, with subtle facial expressions rather than exaggerated acting. 22–27 seconds: She smiles, takes another small bite and turns the camera slightly to show the cozy café interior, Japanese customers naturally sitting in the background and warm practical lighting. 27–30 seconds: She walks back outside onto the Japanese street, looks directly into the phone camera and naturally says, “Okay, this place was actually so good!” She smiles and ends the recording casually. Authentic Japanese environment, realistic human behavior, natural walking physics, real-looking food textures, accurate skin texture, natural hair movement, realistic reflections, imperfect smartphone framing, subtle autofocus hunting, automatic exposure adjustments, slight handheld shake, natural motion blur, realistic low-light smartphone noise, true-to-life colors, detailed background activity, genuine vlog energy, high-resolution smartphone footage, photorealistic, documentary realism, no cinematic camera moves, no drone shots, no artificial slow motion, no CGI appearance, no text, no subtitles.
≡Provenance · verifiedPublished prompts are human-checked against the original post; platform-reconstructed prompts are not accepted.Compare with source ↗

Models & stability

Recommended models

Seedance 2.5
Stability
82 / 100
Cost
Medium
Related SkillsPOV and vlog presence ↗Food close-ups and eating ASMR ↗

Independent retest evidence

See what the same prompt generated in our test.

Retest output · not the creator's original

Retest reading · GoodCase editors

What the creator aimed for
The creator wanted a 30-second Tokyo phone food vlog: talking to camera while walking, finding a small dessert cafe, tasting a strawberry dessert, and a closing line outside, all looking like casual flagship-phone footage.
Actual input and models
A text-only prompt in six timed segments, followed by a long list of phone-look requirements and bans. The creator lists Seedance 2.5. We retested on MiniMax H3 Max 768p, confirmed by the provider receipt.
Test conditions and runs
2026-09-07, MiniMax H3 Max at 768p, one generation. Two AI judges scored it automatically; no human verdict. The length of our output is not recorded in this data.
What held up
The one closing line written in the prompt shows up verbatim in the transcript. The judges gave 84.3 and 80.5, a 3.8-point gap, the smallest among this page's representatives.
What failed or can't be told
The transcript also has an opening line the prompt never wrote; it only said she talks casually, so the model filled in words itself. About 17% of the audio is silent, and the data does not show where. Whether the autofocus and exposure details came through is not visible in the evidence.
Variables to try next
Write the opening line out if you want control over it; keep the six-segment structure and swap the city and food; run it on Seedance 2.5 for comparison; run it several times before judging consistency.

Written by GoodCase from the retest data below and the creator's prompt; it is our reading, not the creator's statement.

VideoGenerated evidenceAI judged

MiniMax H3 Max 768p

Sep 7, 2026

Retest score82.4
Your browser does not support video playback.
claude-sonnet-5 · 84.3gpt-5.6-sol · 80.5gap · 3.8
  • Cross-model retest: the recommended model is Seedance 2.5, and this run used MiniMax H3 Max 768p. The result shows how the same prompt carries over to MiniMax H3 Max 768p; it does not show what the creator got with the original model.
  • The retest score is the mean of two AI judges (0–100) for this single generation. No person has reviewed it, and it is not a success rate.

Model and resolution follow the provider task receipt.

Keep exploring

Explore related cases.

View all AI Video Case →
A Tokyo Girl’s Gym Day in the Early 2000s
AI VideoXAdded 10/05/2026

A Tokyo Girl’s Gym Day in the Early 2000s

A little old school workout session.🏋️‍♀️ Seedance 2.5 on @openart_ai Prompt Create a highly photorealistic 30-second early-2000s Japanese gym day vlog featuri…

Skill · POV and vlog presence

Prompt

A little old school workout session.🏋️‍♀️ Seedance 2.5 on @openart_ai Prompt Create a highly photorealistic 30-second early-2000s Japanese gym day vlog featuring ONE consistent young Japanese wo…

Source heat · X
40
Stability
View case →
Korean Idol VHS Beauty Vlog
AI VideoXAdded 10/03/2026

Korean Idol VHS Beauty Vlog

A MiniMax H3 cinematic narrative prompt supplied by the original author's public X post, with the video result verified.

Skill · Early-2000s DV home videoSkill · POV and vlog presence

Prompt

DV 16mm camcorder vlog with soft VHS tape quality, handheld POV by CHASE and mirror-mounted vanity shots. CHASE, a Korean idol in her 20s with long black hair and dewy skin, casual

Source heat · X
86
Stability
View case →
Tokyo Alley Strawberry Dessert Café Vlog
AI VideoXAdded 10/03/2026

Tokyo Alley Strawberry Dessert Café Vlog

Both MiniMax H3 Max and MiniMax H3 currently offer unlimited generations on @itsPolloAI. So now is a great time to try them out for yourself. Prompt of this vid…

Skill · POV and vlog presence

Prompt

Both MiniMax H3 Max and MiniMax H3 currently offer unlimited generations on @itsPolloAI. So now is a great time to try them out for yourself. Prompt of this video ↓ Realistic smartphone food vlog s…

Source heat · X
48
Stability
View case →

Same creator

More from @saniaspeaks_ →

  • A Korean Girl’s Early-2000s Birthday CelebrationAI Video · Seedance · 2026-10-02
  • Schoolgirl Steals a Diamond Necklace and Escapes by TrainAI Video · Seedance · 2026-09-24
  • A Korean Woman’s Summer Day on a Traditional FarmAI Video · Seedance · 2026-09-17
  • Tokyo Street Racer Girl with a Vintage MotorcycleAI Image · GPT Image · 2026-09-17
  • Editorial Poster Pairing a Photo with Handmade Paper ArtAI Image · GPT Image · 2026-10-01
  • Young Korean Woman Learns to Milk a CowAI Video · Seedance · 2026-09-26

Same model

Latest with Seedance 2.5 →

  • A Woman in Pink Shares Playful Moments with Her Boyfriend in BusanAI Video · Seedance 2.5 · 2026-10-23
  • Two Superfast Women Meet in a City Frozen in TimeAI Video · Seedance 2.5 · 2026-10-23
  • Toddler Feeds Dad Cash as Mom Joins a Playful Toy Money SchemeAI Video · Seedance 2.5 · 2026-10-23
  • A Seoul Girl’s Day from Morning Makeup to City StrollsAI Video · Seedance 2.5 · 2026-10-23
  • A Child’s Day in 2000 Through Their Own EyesAI Video · Seedance 2.5 · 2026-10-23
  • An Elderly Woman’s Red Scarf and First Love Beneath a Maple TreeAI Video · Seedance 2.5 · 2026-10-23

Each detail page keeps the original result, the prompt, and retest evidence in one reading path.