How AI Rumpelstiltskin Works

A straightforward pipeline with no hidden steps: your photo goes in, a playful dance video comes out, and every credit movement is recorded in an auditable ledger.

1. You upload an authorized photo

Sign in with Google, then upload one or two clear images (JPEG or PNG, up to 10 MB each). You confirm you own the image or have permission to use the person's likeness. Photos are stored privately and are never shown publicly.

2. We reserve credits and submit the job

Before you generate, the Studio shows the exact credit cost. When you confirm, that amount is reserved from your balance and a durable job is queued. Nothing is charged if generation is unavailable.

3. The model renders your dance scene

Our video model creates a Rumpelstiltskin-inspired scene from your photo. Rendering is asynchronous: it can take several minutes to tens of minutes depending on queue and model load. We do not promise fixed turnaround times.

4. You preview and download privately

When the job finishes, the video appears in your dashboard. Preview it there, download it while it is retained, or delete it at any time. If the job definitively fails, the reserved credits are released automatically.

What to expect from results

Generative video varies with the input photo and model behavior. Faces are usually recognizable, but artifacts can occur; an exact recreation of any existing viral clip is neither promised nor attempted. Jobs can fail — when they do, credits reserved for that job are released, and your dashboard shows what happened.

Your content, your control

Uploaded originals are kept for a short retention window (currently 72 hours) and then removed; finished videos are kept for 30 days by default, and you can delete them sooner. Everything is private by default — there is no public gallery of user creations.

More questions? See the FAQ, read the Privacy Policy, or start with the photo-to-video tutorial.