GoodCase.aiGoodCase.aiAI Case Evidence Library
RankingsCasesSkillsCreatorsFavoritesSubmitDaily
中文EN
Agent Access →
GoodCase.ai · AI Case Evidence LibraryLoading the case library.
ChangelogFeedback
GoodCase.aiGoodCase.aiAI Case Evidence Library
RankingsCasesSkillsCreatorsFavoritesSubmitDaily
中文EN
Agent Access →
AI VideoX2026-08-03Evidence L1#youmind#Seedance 2.5Seedance 2.5

Bladesmith Competition Reality TV Video

TechHalla avatarCreatorTechHalla →

A realistic reality TV style video prompt showing a bladesmith testing a forged machete against ceramic gnomes with dynamic impact physics and documentary camera work.

View original source ↗
Source ♥ 514

Prompt

Reality TV bladesmith competition testing segment, multi-camera documentary production, bright industrial workshop lighting, sharp HD broadcast look, practical forge practicals, natural handheld shake on action cams, clean documentary texture, no cinematic film look.

0-4s: [Medium Two-Shot] Martial arts master of clear Asian ancestry in black tactical shirt stands facing the young female bladesmith in her early 20s (work apron, focused). He holds her forged machete, turns the blade once under the lights. Looks her in the eye: “Are you ready?” She answers firmly “Yes.” Quick cut to her determined face, then to three classic ceramic garden gnomes lined up on sturdy stands behind them.

4-7s: [Low-Angle Tracking + Rapid Cuts] Master steps in, raises the machete. Hard cut to side angle as he delivers a powerful overhead chop into the first gnome. Impact explodes in slow motion: ceramic shatters into sharp fragments, paint chips and dust fly outward. Time locks at peak penetration. Camera orbits the frozen blade buried in the broken gnome for one second. Quick insert ECU of the edge biting through the ceramic body.

7-11s: [Dynamic Side Tracking] Action resumes. Master yanks the machete free, immediately follows with a heavy diagonal slash across the second gnome. Slow-motion again: the figure tears apart sideways, large colored pieces and shards scatter. Freeze at maximum destruction. Two rapid cuts: one over-shoulder from behind the master, one low angle under the flying fragments. Orbit holds for half a second on the locked impact.

11-15s: [Handheld Close Action] Third strike — short, vicious downward chop into the third gnome. Slow-mo impact, the ceramic erupts into dozens of pieces, hat and beard flying. Full freeze. Camera pushes in and circles the frozen chaos for one beat. Quick insert of the master’s focused face mid-swing, then back to the locked blade.

15-19s: [Series of Fast Cuts] Time resumes. Master pulls the machete out, turns it under the light. Tight ECU tracks along the edge: a few small nicks and rolled spots visible. Cut to his fingers testing the edge. Cut to the ruined garden gnomes still shedding fragments. Cut to the young woman’s tense expression as she watches.

19-24s: [Medium Close-Up] Master lowers the blade, looks straight at her. Calm, professional delivery: “There are a couple of nicks on the edge… but your weapon, ma’am… it kills.” Hard cut to her face — a small, controlled smile of pride breaks through.

24-28s: [Final ECU Montage] Extreme close-up of the nicked but still lethal edge catching the light. Cut to the shattered remains of the garden gnomes. Final frame holds on the young bladesmith’s quiet satisfaction as the master nods once.

Realistic TV broadcast quality, detailed ceramic shatter physics, sharp edge detail, natural motion blur only on moving elements, stable characters, coherent impact timing, no artifacts, high rewatch value.

Reusable method

Key decisions

Check whether camera continuity, transition logic, and action resolution all hold in the final result.

Adapt it to your subject

Extract “target result → the subject, scene, camera movement, and final action → constraints” as a reusable template for similar work.

Where it breaks

Run single-variable comparisons around shot length, transitions, and motion amplitude. Record the model, cost, and failures; turn it into a Skill only after it repeats reliably.

Recommended models

Seedance 2.5
Stability
Cost
Medium
This case has not yet been distilled into a Skill.Request it in feedback →
TechHalla avatar

Creator

TechHalla

TechHalla focuses on AI video production and is active on X. “Frozen Time and Rewind in a 1950s Diner” is the current representative case, built with Seedance 2.5 with its stability retest open to votes.

Editor’s pickAI VideoHigh source heat
View creator →

Keep exploring

Explore related cases.

View all AI Video Case →
AI VideoX08/07/2026MAX →

16mm Analog Morning Vlog

A highly detailed video prompt for Seedance 2.5 mimicking a handheld 16mm analog morning vlog of a woman walking her dog, focusing on realistic imperfections and film texture.

Skill · POV and vlog presence

Prompt

CAMERA: Handheld DV 16mm daily vlog footage. The video MUST begin with her holding the camera at arm's length in selfie mode while stepping outside her apartment building with her

Source heat
81
Stability
View case →
AI VideoX08/06/2026TechHalla →

90s Diner Time Freeze Effect

A cinematic video prompt for Seedance 2.5 featuring a high-physics diner collision where time freezes, allowing a character to interact with suspended objects before resuming in real-time.

Prompt

Photorealistic cinematic 1990s American diner, red vinyl booths, neon signs, chrome details, checkerboard floor, soft natural window light mixed with warm practicals, subtle handhe

Source heat
83
Stability
View case →
AI VideoX08/04/2026John →

Bamboo Toy Car Transformation Vlog

An incredibly detailed action prompt for a 30-second handheld POV video featuring a girl with a bamboo toy that triggers a car transforming into a giant robot.

Skill · POV and vlog presence

Prompt

纯属虚构的荒诞动作短片。100%真实手机实拍质感,iPhone纪录片风格,30秒手持一镜到底,无剪辑、无转场。自然日光,轻微呼吸式镜头运动,偶发自动对焦迟疑,轻微滚动快门,真实动态模糊,镜头边缘带少量指纹和灰尘。所有变形、机械运动、撞击和碎片效果都必须具有沉重的现实物理反馈,呈现为全尺寸机械装置与实景特效,绝不能出现电子游戏、动画或廉价CG质感。保留现场原始

Source heat
67
Stability
View case →
GoodCase.ai · AI Case Evidence LibraryEach detail page keeps the work, method, creator, and retest evidence in one reading path.
ChangelogFeedback