Current generation model
Seedance 2.5 through RunningHub generates one continuous silent video from the fixed performance reference. Uploaded photos first become hand-drawn colored-pencil portraits through fal’s GPT Image 2 API. fal handles video-only clipping when needed; our private server restores the existing soundtrack and checks the final MP4. These providers do not endorse this service. Models from earlier private experiments are not the current generation route.
Cloud media processing
fal provides portrait editing, temporary media storage and video-only clipping. Our own private server adds the saved soundtrack and checks MP4 timing and format. These checks do not themselves establish that a result is safe or that the model preserved an identity or singing mouth motion.
Reference photos and performances
Photos provide identity and appearance references. For a 19-second request, the fixed site template guides staging, motion, camera and timing. The workflow generates two segments and assembles one file. This is a reference-conditioned generation process, not a guarantee of pixel-identical source editing.
Audio and rights
No original commercial song or third-party performance is bundled with your purchase. The fixed template’s soundtrack is applied automatically. Users do not upload or select audio. The operator must confirm rights in the template and soundtrack before public rendering is enabled. Keeping an audio track does not prove that mouths follow every word.
Known quality limitations
The tested workflow can preserve a broad band arrangement, but faces, clothes, backgrounds, gestures, hands and instruments can change. A role may be assigned incorrectly in a close-up, and mouthpiece contact or lip sync can be inaccurate. Exact reproduction, perfect likeness, a precise musical performance and unrestricted commercial rights are not promised. Confirmed material defects remain subject to the refund policy.
AI disclosure and examples
Results are synthetic creative media, not recordings of an actual performance. Describe them as AI-generated when sharing and follow the destination platform’s labeling rules. Product pages distinguish third-party trend references from our own samples. We do not currently claim a visible watermark or embedded provenance mark on every generated file.
Availability and safety
Saving a draft is free and does not place a paid order. When checkout is available, Waffo confirms the one-time payment before generation starts. Return to the original browser or use the purchase email in Order lookup to claim an eligible paid video; no verification email is sent. Inputs and outputs require review, and restricted uses remain prohibited even if a model can technically produce them. Publishing model information is not a statement of payment-provider approval.