ChatGPT got me a fake cooking video worth 2.8M views and $2.5K at TikTok's rate

No stove, no ingredients, no camera, no real dish. Two prompts in, a 20-second "recipe video" out.
Full pipeline below - prompts, tools, CapCut and every platform that pays for AI video in 2026.
THE ONE-LINE VERSION: AI generates your dish photo - AI generates 10 scene images from it - each image animates into a clip - stitch + subtitle in CapCut - upload - get paid.
Two finished outputs from this exact pipeline, back to back - this is what "no filming" actually looks like.
STEP 1 - THE DISH (no photo needed)
Skip the camera and skip stealing a photo off Pinterest - both create problems later. Instead you generate your own reference photo from scratch with one prompt.
COPY THIS PART AS-IS ↓
YOU DECIDE:
[YOUR DISH] - e.g. "creamy garlic chicken with spinach, sun-dried
tomatoes, and parmesan"
This generated image becomes your reference for every scene in Step 2 - no real dish, no real photo, ever existed.
STEP 2 - THE SCENE PROMPT
Don't ask for one image - ask for 10 numbered scenes that "cook" the dish. Same skeleton every time:
Aspect ratio 9:16 + SCENE [N] - [ACTION] + Angle Blueprint + who does what + camera/motion blur/steam.
Faceless hands only, natural light, motion blur on movement - that sells "real footage."
THE UNIVERSAL PROMPT:
COPY THIS PART AS-IS ↓
YOU DECIDE THESE (fill in once, reuse everywhere they appear below) ↓
COPY THIS PART AS-IS, brackets auto-fill from your list above ↓
Send this in the same chat where you generated the Step 1 image - no attachment needed, it's already the reference.
Reality check: the batch "prompt queue" (video above) runs on a tool built on the image API, not plain chat. In vanilla ChatGPT you'll get 2-3 images per turn - nudge it with "continue with scene 4."
STEP 3 - IMAGE TO VIDEO
Drop each scene image into an image-to-video generator (Higgsfield/Seedance, Kling, Runway):
10 scenes - 10 clips - ~20s of "footage."
STEP 4 - CAPCUT
Timeline in scene order - trending audio - cut tight to the beat - word-by-word subtitles (half the retention).
Mobile: Text - Auto captions - pick audio source - Create.
Desktop/web: Text/Captions tab - Auto captions - choose language - Create - style with AI presets.
Export 9:16, 1080p.
STEP 5 - WHERE TO GET PAID (2026)
TikTok Creator Rewards - 10K followers + 100K views/30d, 60s+ videos, $0.40-1.20/1K views.
YouTube Shorts - 1K subs + 10M views/90d for ad share, $0.03-0.13/1K.
Facebook Content Monetization - merged program, ~5-10K followers, $0.02-0.15/1K.
Instagram Reels - traffic/affiliate channel, bonuses mostly gone.
X Revenue Sharing - Premium + 500 verified followers + 5M impressions/90d, ~$8-12/million impressions.
Selling the pipeline - restaurants/food brands need this, don't want to film. $200-1,500/video.
Affiliate - kitchenware/ingredient links in bio, no follower threshold.
Label AI content - TikTok/YouTube/Meta pull unlabeled or reused AI clips from monetization.
Thanks for reading this far - now go make your own version.
Then never touch a stove, a camera or Pinterest again)
Prompts
Animate this image, 9:16, seamless. [the motion - pours/sprinkles/
tosses/lifts]. Subtle handheld shake, motion blur, steam rising.
Duration 2s.Using the image you just generated as the reference, create a 10-scene
image set, aspect ratio 9:16 each. This is [YOUR DISH]. Generate one
photorealistic image per scene, faceless hands only, natural kitchen
lighting, motion blur on moving ingredients, steam where relevant.
[YOUR DISH] - e.g. "spicy beef tacos with lime and cilantro"
[MAIN INGREDIENT] - e.g. beef
[SECONDARY INGREDIENT 1 / 2] - e.g. onions, peppers
[SAUCE OR LIQUID BASE] - e.g. salsa
[GARNISH 1 / 2] - e.g. lime, cilantro
[VESSEL / COOKING VESSEL] - e.g. skillet, pan, pot
[CONTAINER] - e.g. jug, bottle, spoon
[UTENSIL] - e.g. fork, spoon
[DONENESS CUE] - e.g. golden, bubbling, meltedSCENE 1 - HOOK / FINISHED [YOUR DISH] MONEY SHOT
Angle Blueprint: ANGLE A - TOP-DOWN HERO SHOT
The finished [YOUR DISH] in a [VESSEL], garnished with [GARNISH 1] and
[GARNISH 2], steam rising. Camera directly overhead, dish fills 80% of
frame, soft natural light, [SAUCE OR TEXTURE] as the hero.
SCENE 2 - SETUP / INGREDIENT LAYOUT
Angle Blueprint: ANGLE B - FLAT LAY
Raw ingredients on a [SURFACE]: [MAIN INGREDIENT], [SECONDARY
INGREDIENT 1], [SECONDARY INGREDIENT 2], [GARNISH 1], [GARNISH 2],
[SAUCE OR LIQUID BASE], in small bowls. Camera overhead, symmetrical.
SCENE 3 - ADD [MAIN INGREDIENT] TO [COOKING VESSEL]
Angle Blueprint: ANGLE C - OVERHEAD MATERIAL SHOT
Faceless hands placing raw [MAIN INGREDIENT] into a [COOKING VESSEL]
on the stove. Camera high angled, mid-motion with motion blur, oil
glistening.
SCENE 4 - COOK [MAIN INGREDIENT] UNTIL [DONENESS CUE]
Angle Blueprint: ANGLE E - COOKING / HEATING DEMO SHOT
A faceless person cooking [MAIN INGREDIENT] in a [COOKING VESSEL]
until [DONENESS CUE]. Camera 35-degree downward front angle, one hand
on the handle, the other working the food. Visible cooking detail.
SCENE 5 - ADD [SECONDARY INGREDIENT 1]
Angle Blueprint: ANGLE C - OVERHEAD MATERIAL SHOT
One hand tipping [SECONDARY INGREDIENT 1] onto the [MAIN INGREDIENT].
Camera high angled, mid-motion with motion blur.
SCENE 6 - ADD [SAUCE OR LIQUID BASE]
Angle Blueprint: ANGLE J - TRACKING POUR SHOT
[SAUCE OR LIQUID BASE] poured from a [CONTAINER] into the [COOKING
VESSEL]. Camera tracking the pour, mid-splash motion blur, steam
rising.
SCENE 7 - ADD [SECONDARY INGREDIENT 2]
Angle Blueprint: ANGLE C - OVERHEAD MATERIAL SHOT
One hand tipping [SECONDARY INGREDIENT 2] into the [COOKING VESSEL],
falling with motion blur. [MAIN INGREDIENT] visible beneath.
SCENE 8 - ADD [GARNISH 1]
Angle Blueprint: ANGLE C - OVERHEAD MATERIAL SHOT
One hand scattering [GARNISH 1] over the [YOUR DISH]. Camera high
angled, mid-toss motion blur.
SCENE 9 - ADD [GARNISH 2]
Angle Blueprint: ANGLE H - GARNISH SHOT
One hand sprinkling [GARNISH 2] over the finished [YOUR DISH]. Camera
medium-high angle, slight motion blur.
SCENE 10 - LIFT / TEXTURE REVEAL
Angle Blueprint: ANGLE F - LIFT / TEXTURE REVEAL SHOT
A [UTENSIL] lifting a portion of [YOUR DISH], showing texture. Camera
close, front angle, plate below defocused, texture as the hero.
Output all 10 as a single set.Generate one photorealistic image of [YOUR DISH], fully plated and
finished, aspect ratio 9:16, natural kitchen lighting, top-down angle,
steam rising, glossy and appetizing. No text, no hands, no logos.Related articles

ChatGPT got me a fake cooking video worth 2.8M views and $2.5K at TikTok's rate
No stove, no ingredients, no camera, no real dish. Two prompts in, a 20-second "recipe video" out.

YouTube's biggest channels don't always make good videos..?
A lot of people assume that YouTube views and subscribers are an indication of how good the videos are.

Claude Helped me get 588,000 Views With Instagram Reels AI Can Cite (9-Step Playbook)
I generated 588,000 views from Instagram Reels by spending about 15 minutes per day. And I only started posting 60 days ago..