🔩 Rust Removal Videos — Complete Master Guide
Create cinematic rust-to-restored transformation content using Stable Diffusion and Google Flow — 4 image prompts in JSON format plus 3 video prompts for each antique object, complete with a second act where an elderly figure carries the restored object back to the world it came from.
What Is This Niche?
These videos follow a two-act cinematic structure. Act one: a rusted antique object — sword, lantern, compass, telescope, padlock — sits on a surface. An aged hand enters, drenches it in chemical spray, foam erupts violently from every rust contact point simultaneously, and a single cloth wipe reveals the fully restored object beneath. Act two: the elderly figure carries the restored object back to the world it came from — a foggy forest, an abandoned lighthouse, a stone dungeon gate — and holds it in a final dramatic pose at the place where it was forgotten.
💡 The Core Hook: Two of the most powerful viral mechanics combined in one format — the satisfaction of a transformation reveal and the emotional weight of returning something lost back to where it belongs. The foam eruption that hides the object completely before the single-wipe reveal is the most-watched 3 seconds in this entire niche.
Why Does This Go Viral?
| Element | Why It Works |
|---|---|
| 🧼 Foam Eruption Reveal | The moment all foam erupts simultaneously from every surface at once is the most physically satisfying visual event in short-form content — physically impossible to scroll past |
| 🧽 Single Cloth Wipe | One continuous uninterrupted wipe revealing the restored object beneath is the highest-save moment — viewers watch the full stroke every replay |
| 🦾 Aged Hands | Deeply wrinkled, veined, calloused hands create instant emotional authenticity — viewers feel the weight of the object through those hands |
| 🏰 Environment Matching | The object's environment is chosen for historical and emotional resonance — a sword belongs in fog and ruins, a compass at a sea cliff — this storytelling layer drives saves and shares |
| 🔬 Macro Detail | Individual rust flakes, pitting holes, seized mechanisms, verdigris bloom — 8K macro detail at the level of still photography triggers deep visual fascination |
| 🔄 Before and After | The same object, same surface, same angle — only the condition changed. The brain rewards this comparison with dopamine. Proven highest-retention format on all platforms. |
| 🌍 No Narration, No Music | Only diegetic sounds — spray hiss, foam crackling, cloth wipe, footsteps — universal content that works in every language and culture |
| ♾️ Unlimited Object Library | Any rusted iron, brass, steel, or copper antique generates a completely unique 4-image + 3-video sequence — the format never repeats itself |
Complete Output — 4 Images + 3 Videos Per Object
Each object generates a full production package. Every prompt is production-ready — images in Stable Diffusion JSON format with all parameters locked, videos in Google Flow format with settings blocks.
Image 01
Rusted Object on Surface
Macro close-up, every rust detail labeled in CAPS sections. Rust liquid seeping onto surface beneath. No people.
Image 02
Restored Object on Surface
Identical angle, identical surface, identical lighting. Only the object condition changed. Faint warm reflection on surface.
Image 03
Elderly Male — OTS Walk
Over-shoulder POV. Figure carrying restored object toward the historically matched destination environment ahead.
Image 04
Elderly Male — Final Pose
Low angle ground-level shot looking upward. Figure and restored object monumental against the environment background.
Video 01
Rust Removal Transformation
Spray → foam eruption → single wipe reveal. Camera locked, no cuts, 15 seconds. The core viral moment.
Video 02
Pick Up and Travel
Hand lifts object → OTS walk through environment → figure arrives at destination. 15 seconds across cuts.
Video 03
Dramatic Action at Destination
Figure performs the final meaningful action at the historically correct environment — sword raised in fog, lantern lifted at ship bow, telescope pressed to eye at lighthouse. 15 seconds. Final stillness.
Video Structure — 3 Acts Across 3 Videos
Video 01
The Transformation — Spray, Foam, Wipe
Camera locked in macro close-up. Aged hand enters with spray bottle and drenches every surface without exception — liquid running down all sides carrying rust residue onto the surface beneath. One second pause. Foam erupts from every rust contact point simultaneously with violent sudden force — builds until the object is 100% invisible. Four seconds of stillness. One single cloth wipe from end to end reveals the fully restored object as the foam peels back like a curtain. Hand withdraws. Object alone. Complete silence.
Video 02
The Journey — Lift, Walk, Arrive
Macro on restored object. Aged hand descends and lifts it with visible care — the object catches available light as it rises. Cut to over-shoulder POV walking through the historically matched environment toward the destination. Cut to the figure stopping at the destination, coat settling, object at side. The environment surrounds — fog, torchlight, ocean night, coastal wind. The figure stands completely still.
Video 03
The Moment — Action, Hands, Final Stillness
Figure moves through the environment toward the final action point. Close-up on aged hands performing the key action — gripping, raising, pressing to eye, placing at the gate. Low angle wide shot as the figure completes the final action and holds completely still. Only ambient movement — fog drifting, torch flame, ocean wind. The figure does not move. The object is where it belongs.
Tools You Need
- Claude or ChatGPTPaste the master prompt — receive 10 antique object options. Pick one and receive the complete 4-image JSON package and 3-video Google Flow package in two separate sections
- Stable Diffusion (realisticVisionV60B1_v51VAE)All 4 image prompts are pre-formatted as Stable Diffusion JSON with model, sampler, seed, LoRAs, and HiRes fix — paste directly into your SD interface
- Google Flow (Veo 3)All 3 video prompts are formatted for Google Flow with aspect ratio, duration, camera motion, start frame, end frame, and seed — paste directly
- CapCutJoin the 3 video segments, add diegetic sounds (spray hiss, foam crackling, cloth wipe, footsteps, ambient environment), export 9:16 vertical — no music, no color grading
Generation Strategy
Step 1 — Get 10 Object Options and Pick One
Paste the master prompt into Claude or ChatGPT. The AI immediately outputs a numbered list of 10 rusted antique objects — weapons, navigation instruments, maritime hardware, tools, locks, household items — without waiting for instruction. After the list it asks: "Pick a number 1–10 and give me your surface, environment, and character — or leave blank for defaults." Pick a number. You can specify a custom surface or environment, or leave blank to let the prompt choose the historically correct environment for that object automatically.
Step 2 — Receive 4 Image Prompts in Stable Diffusion JSON
The AI outputs all 4 image prompts as a complete block in Stable Diffusion JSON format — model locked to realisticVisionV60B1_v51VAE, sampler DPM++ 2M Karras, all 4 prompts sharing the same seed, LoRAs add_detail and epi_noiseoffset pre-set, HiRes fix enabled with 4x-UltraSharp. Every object part is described in separate labeled CAPS sections. Paste each JSON block directly into your SD interface with no editing required.
Step 3 — Receive 3 Video Prompts for Google Flow
Immediately below the image block, the AI outputs all 3 video prompts as separate copy-paste blocks for Google Flow — each with a settings block showing aspect ratio (9:16), duration (15 seconds), camera motion description, start frame, end frame, and seed. Use the generated Images 01 and 02 as start and end frames for Video 01. Videos 02 and 03 use the character and environment images as references.
Copy the Master Prompt
Paste this entire prompt into Claude or ChatGPT. The AI immediately outputs 10 antique object options with no preamble. Pick a number and receive your complete production package — 4 Stable Diffusion JSON image prompts + 3 Google Flow video prompts, all sharing a locked seed, all ready to paste directly into your tools.
master-prompt.txt
You are a cinematic AI image and video prompt specialist for a rusted object restoration content series. ════════════════════════════════════════ FIRST ACTION — do this immediately and automatically: ════════════════════════════════════════ The moment you receive this prompt, before anything else, immediately generate a numbered list of exactly 10 rusted aged or weathered objects. Do not wait for any instruction. Do not say anything before the list. Just output the list immediately. List rules: — Physical iron, brass, steel, copper, or mixed-metal objects only — Vintage, antique, or industrial era objects only — Objects that can realistically rust, corrode, or oxidize...
You are a cinematic AI image and video prompt specialist for a rusted
object restoration content series.
════════════════════════════════════════
FIRST ACTION — do this immediately and automatically:
════════════════════════════════════════
The moment you receive this prompt, before anything else, immediately
generate a numbered list of exactly 10 rusted aged or weathered objects.
Do not wait for any instruction. Do not say anything before the list.
Just output the list immediately.
List rules:
— Physical iron, brass, steel, copper, or mixed-metal objects only
— Vintage, antique, or industrial era objects only
— Objects that can realistically rust, corrode, or oxidize
— Strong visual restoration transformation potential
— Mix categories: weapons, tools, navigation instruments, household
items, maritime objects, mechanical devices, locks and hardware
— No fantasy objects. No modern objects. No plastic or wooden objects.
— Each object visually distinct from the others
— Numbered list only. No descriptions. No explanations.
After the list, say exactly this one line:
"Pick a number 1-10 and give me your surface, environment, and
character — or leave blank for defaults."
Then wait for my selection.
════════════════════════════════════════
When I give you a number or object name — generate exactly:
— 4 IMAGE PROMPTS (Stable Diffusion JSON format)
— 3 VIDEO PROMPTS (Google Flow format)
IMAGE PROMPTS and VIDEO PROMPTS must be in two completely separate
sections. Never mix them.
Generate ALL 4 image prompts first as one complete block.
Then generate ALL 3 video prompts as separate individual copy-paste
blocks with settings below each one.
════════════════════════════════════════
IMAGE PROMPT RULES — all 4 prompts:
════════════════════════════════════════
IMAGE 01 — Rusted object on surface (no people)
IMAGE 02 — Restored object on same surface (no people, identical setting)
IMAGE 03 — Elderly male in worn coat holding restored object approaching
destination — over shoulder POV
IMAGE 04 — Elderly male with object in dramatic pose at destination —
low angle upward shot
Surface for IMAGE 01 and 02:
Use the surface I specify. Default: matte black surface.
Environment for IMAGE 03 and 04:
Use the environment I specify. Default: dark stone corridor torch light.
All 4 image prompts must use this exact JSON format:
{
"prompt": "...",
"negative_prompt": "...",
"model": "realisticVisionV60B1_v51VAE",
"width": 832,
"height": 1472,
"steps": 42,
"cfg_scale": 7.5,
"sampler": "DPM++ 2M Karras",
"clip_skip": 2,
"seed": [same seed all 4 prompts — pick one random seed and lock it],
"loras": [
{"name": "add_detail", "strength": 0.85},
{"name": "epi_noiseoffset", "strength": 0.6}
],
"hires_fix": {
"enabled": true,
"upscaler": "4x-UltraSharp",
"upscale_by": 1.5,
"denoising_strength": 0.3
}
}
IMAGE 01 prompt rules:
— Extreme close-up macro photography
— Every part of object described in separate labeled sections in CAPS
— Deep corrosion, thick flaking rust, pitting, seized mechanisms,
verdigris bloom, rust streaks, rust dust, zero metallic shine
— Surface detail: rust stain bleeding onto surface beneath object
— Rust liquid seeping from object base onto surface
— Camera: macro 100mm, f/2.8, shallow DOF, 8K RAW no color grading
IMAGE 02 prompt rules:
— Identical shot angle as IMAGE 01
— Identical surface as IMAGE 01
— Identical lighting as IMAGE 01
— Only the object condition changes — fully restored
— Satin finish, crisp markings, clean mechanisms, warm material tones
— Faint warm reflection from restored object glowing on surface beneath
— Same labeled sections as IMAGE 01 describing restored condition
— Camera: macro 100mm, f/2.8, shallow DOF, 8K RAW no color grading
IMAGE 03 prompt rules:
— Over-the-shoulder POV
— Shoulder visible bottom right foreground in soft bokeh
— Elderly male, heavy worn dark wool coat, aged weathered hands
— Restored object carried at waist height catching available light
— Destination environment visible ahead in frame
— Camera: 35mm, f/2.0, natural available light, RAW no color grading
IMAGE 04 prompt rules:
— Low angle ground level shot tilted upward 25-30 degrees
— Elderly male in dramatic pose — object raised or presented
— Figure and object monumental against environment background
— Camera: 24mm wide angle, f/2.8, deep DOF, RAW no color grading
════════════════════════════════════════
VIDEO PROMPT RULES — all 3 prompts:
════════════════════════════════════════
VIDEO 01 — Rust removal transformation
Start Frame: Rusted object on surface (IMAGE 01)
End Frame: Restored object on surface (IMAGE 02)
Mandatory sequence — follow exactly:
STEP 1 — Camera locked. Zero movement. No cuts throughout entire video.
STEP 2 — Aged elderly hand enters frame. Deeply wrinkled skin. Prominent
veins. Age spots. Calloused fingertips. Slight natural tremor. Holding
large chemical spray bottle.
STEP 3 — SPRAY SEQUENCE: multiple slow passes from all angles — top pass,
side passes, underneath where reachable — every surface of the object
completely drenched without exception — liquid running down all surfaces
carrying rust residue — pooling in every crevice and mechanism gap —
rust rivulets running from object base onto surface beneath — not a
single dry surface remaining — entire object glistening wet — hand
withdraws fully from frame.
STEP 4 — One second pause. Complete stillness.
STEP 5 — FOAM ERUPTION: foam erupts from every single surface
simultaneously at the exact same moment with violent sudden force —
every rust contact point bursting at once — foam expands rapidly and
aggressively from all sides equally — foam builds consuming the entire
object completely — within 4-5 seconds the object is 100% invisible
under foam — no part of object visible — only foam mound — object shape
only suggested by foam mound contour — rust bleeds upward into foam base
turning it deep rust-orange and brown — rust liquid seeps from foam base
onto surface in rivulets.
STEP 6 — Four seconds of stillness. Foam sits heavy. Bubbles pop slowly.
Rust liquid continues seeping onto surface.
STEP 7 — ONE SINGLE WIPE: aged hand re-enters holding large clean white
cloth — one single continuous stroke from one end of object to the other
without lifting — entire foam mass dragged away in one motion like a
curtain being pulled — restored object revealed progressively from one
end to the other — cloth exits frame carrying all foam and rust —
restored object alone on surface. Complete silence.
Sound: spray hiss — foam eruption crackling — bubbles popping — rust
liquid seeping — single cloth wipe — final silence.
15 seconds. No cuts. No color grading. No music.
VIDEO 02 — Pick up and travel to environment
Start Frame: Restored object on surface (IMAGE 02)
End Frame: Elderly male standing at destination holding object
Mandatory sequence:
SHOT 1 — Locked macro on restored object. Aged elderly hand descends
slowly. Same hand description — deeply wrinkled, veins, age spots,
calloused, tremor, worn coat sleeve. Object lifted with care and visible
weight. Object catches available light as it rises. Hand and object
exit frame.
CUT — SHOT 2 — Over shoulder POV walking through environment toward
destination. Object carried at waist height swaying with each step.
Environment details passing on both sides. Destination visible ahead
growing larger with each step.
CUT — SHOT 3 — Figure arrives at destination and stops. Coat settles.
Figure completely still. Object at waist height. Environment surrounds.
Stillness.
Sound: object lifting from surface — footsteps in environment — ambient
environment sounds — footsteps slowing and stopping — final stillness.
15 seconds across cuts. No color grading. No music.
VIDEO 03 — Dramatic action at destination
Start Frame: Elderly male at destination holding object (VIDEO 02 end)
End Frame: Elderly male in final pose — object at final position
Mandatory sequence:
SHOT 1 — Opens on figure at destination. Medium or wide shot.
CUT — SHOT 2 — Medium or close-up — figure moves toward final action
point. Deliberate movement.
CUT — SHOT 3 — Close-up on aged hands performing key action with object.
Deep wrinkles. Prominent veins. Age spots. Effort implied in grip.
CUT — SHOT 4 — Low angle wide shot. Figure completes the final action.
Object at final position. Figure holds completely still. Environment
fills background. Only ambient movement — wind, fog, torch flicker,
waves etc. Figure does not move.
Sound: movement sounds approaching action — key action sound specific to
object — final silence with ambient environment sound only.
15 seconds across cuts. No color grading. No music.
Each VIDEO PROMPT must be a separate copy-paste block with settings
block immediately below in this format:
[Video prompt text]
Aspect Ratio : 9:16
Duration : 15 seconds
Camera Motion : [describe]
Start Frame : [describe]
End Frame : [describe]
Seed : [same seed as image prompts]
════════════════════════════════════════
OUTPUT STRUCTURE — always follow exactly:
════════════════════════════════════════
## IMAGE PROMPTS
### IMAGE 01 — Rusted [Object]
[JSON]
### IMAGE 02 — Restored [Object]
[JSON]
### IMAGE 03 — Elderly Male Carrying [Object] OTS
[JSON]
### IMAGE 04 — Elderly Male with [Object] Low Angle
[JSON]
---
## VIDEO PROMPTS
### VIDEO 01 — Rust Removal Transformation
[prompt block]
[settings block]
### VIDEO 02 — Pick Up and Walk to [Environment]
[prompt block]
[settings block]
### VIDEO 03 — [Action] at [Environment]
[prompt block]
[settings block]
════════════════════════════════════════
DEFAULTS if not specified:
════════════════════════════════════════
Surface : Matte black surface
Environment : Dark stone corridor with torch lighting
Character : Elderly male — heavy worn dark wool coat
Aspect Ratio : 9:16 — 832 x 1472
Video Length : 15 seconds each
════════════════════════════════════════
ENVIRONMENT MATCHING RULES:
════════════════════════════════════════
If I do not specify an environment, do not use a default environment
for every object. Instead, choose the environment that best fits the
object's historical context, function, and emotional tone.
A sword belongs in fog and ruins. A telescope belongs at a lighthouse
or sea cliff. A padlock belongs at a dungeon gate. A lantern belongs
on a ship deck at night. A compass belongs at a foggy coastal cliff.
A pocket watch belongs in an overgrown manor at dusk. A hammer belongs
at an abandoned ironworks. An anchor belongs on a derelict stone pier.
Think about what world this object came from. Put it back there.
The environment must feel like the natural resting place of the object
— the place where it was last used, where it was forgotten, where it
belongs. The character carries the restored object back to that world.
Never use the same environment for objects from different worlds.
How To Use — Step by Step
- Copy & Paste the Master PromptCopy the full prompt and paste into Claude or ChatGPT. The AI immediately outputs a numbered list of 10 antique objects — no preamble, no explanation, just the list. Then it asks for your surface, environment, and character — or to leave blank for intelligent defaults.
- Pick a Number — With or Without Custom SettingsType your number (e.g. "4") to use auto-matched defaults, or customize: "4 — slate stone surface, abandoned shipyard, elderly male". If you leave the environment blank, the AI chooses the historically correct world for that object automatically — sword gets fog and ruins, compass gets coastal cliff, padlock gets dungeon gate.
- Generate All 4 Images in Stable DiffusionCopy each JSON block → paste into your Stable Diffusion interface → generate. All 4 images share the same seed for consistency. The model (realisticVisionV60B1_v51VAE), LoRAs, HiRes fix, and all parameters are pre-set — no configuration needed.
- Generate Video 01 — TransformationOpen Google Flow (Veo 3) → upload Image 01 as Start Frame and Image 02 as End Frame → paste Video Prompt 01 with its settings block → generate. This is the core viral video — spray, foam eruption, single wipe reveal. Camera locked, no cuts, 15 seconds.
- Generate Video 02 — JourneyUse Image 02 as Start Frame → paste Video Prompt 02 → generate. The aged hand lifts the restored object, figure walks through the environment toward the destination, arrives and stops. 15 seconds across cuts.
- Generate Video 03 — Dramatic Final ActionUse the Video 02 end frame as reference → paste Video Prompt 03 → generate. Figure moves to the final action point, aged hands perform the key action, low angle shot as the figure holds completely still with the object at its final position. Only ambient movement. 15 seconds.
- Join, Add Sound & ExportImport all 3 videos into CapCut → join in sequence → layer in diegetic sound effects: spray hiss, foam eruption crackling, single cloth wipe, footsteps specific to the environment, ambient environment sounds. No music, no color grading → export in 9:16 vertical format.
- Upload & RepeatUpload to TikTok, Instagram Reels, or YouTube Shorts → paste master prompt again → pick a new object number → a completely fresh 4-image + 3-video production package every single time.
Comments
Post a Comment