Meme Prompts – Rumpelstiltskin AI
Updated
A Rumplestiltskin AI prompt needs an identity image and a complete motion reference. The text alone cannot reproduce a specific sequence: the reference carries its camera, cuts and footsteps.
Copy-ready prompts
1. Solo — one-photo recreation
One reference photo on the lead dancer.
[Reference-to-video · single identity recreation] Motion ref: @Video1 (camera path, blocking, edit rhythm) Identity ref: @Image1 → primary subject Identity lock: high · Temporal consistency: strict across cuts Preserve: scene composition, practical lighting, shot cadence Look: 35mm vintage, soft grain, warm low-key grade Negative: extra subjects, face drift, morphing, flicker, text, watermark
2. Two people — dual identity mapping
@Image1 and @Image2 mapped to the two characters.
[Reference-to-video · dual identity mapping] Motion ref: @Video1 Subject A → @Image1 (lead dancer) · Subject B → @Image2 (seated character) Role mapping: fixed per shot · No identity bleed between subjects Preserve: blocking, framing, edit rhythm, set and grade Look: filmic grain, warm practical key light Negative: identity mixing, duplicated faces, morphing, flicker, watermark
3. “Tiptoe in my Jordans” variant
Solo recreation with a footwear override.
[Reference-to-video · identity + wardrobe accent] Motion ref: @Video1 · Identity ref: @Image1 → lead dancer Wardrobe override: footwear → high-top basketball sneakers (feature in the foot insert) Preserve: choreography, camera path, timing Look: vintage film texture, warm key light Negative: extra limbs, foot artifacts, flicker, text, watermark
Bind each role explicitly
Identify the little dancer by his actions instead of left or right screen position. In a two-person scene, bind each identity image to its own role. Token spelling varies by provider, so check the model’s reference syntax.
Keep the motion separate from identity
A monochrome motion reference controls timing and geometry rather than the final color grade. Faces come from the supplied photos; the scene, costumes and relative character sizes come from the reference instructions.
Keep the original audio in post-production
Generate visuals and restore the reference audio without inventing a replacement soundtrack. A silent generation still needs an authorized source track and suitable synchronization.
Skip the setup
Just me uses one photo for the little dancer and preserves the seated woman. Two of us uses photo one for the dancer and photo two for the woman. Both modes keep the scene costumes, relative character sizes and the reference choreography. The standard file is a 13-second, 720p landscape MP4 (16:9). The original soundtrack is restored after rendering. Video links remain available for 7 days; save the file before the link expires.
Sources
Checked 2026-10-09. External prices and product rules can change.
Frequently asked questions
Can I use two photos?+
Assign the first adult photo to the dancer and the second to the seated woman in Two of us. The shared scene uses 10 credits, just like solo. Watch both faces across the cuts to judge likeness.
What should the reference photo look like?+
Keep a single adult face clearly visible, with soft front lighting and no blur or heavy filters. JPG, PNG and WebP uploads can be up to 8 MB and should be at least 512 pixels on the short side. Both roles follow the scene costumes and relative scale. No celebrities or minors.
Is the Rumplestiltskin AI generator free?+
Rumpelstiltskin AI examples and prompts are free to explore. Selecting a photo gives a local preview; your own moving scene uses a paid pack.
Does it keep the Tiptoe in My Jordans soundtrack?+
Post-production pairs the generated picture with the reference track. Listen to the complete sample: reposts can use different sound. Check your destination platform’s posting rules.
Skip the prompt. Get the video.
No reference clip, tokens or retries to manage — upload one photo and RumplestiltskinAI.com delivers the finished 13-second MP4 with the original sound.
Generate it in one click