Prompt Port Desk

A prompt written for Veo and the same prompt written for Seedance are not one text with different labels on it. They are different shapes over different capabilities, so a port is a lossy transformation rather than a rewrite - Veo to Seedance loses native multi-speaker audio, Kling's fifteen seconds to Veo's eight loses seven of them, Sora's twenty-five to Hailuo's ten loses fifteen. That delta is arithmetic over two capability rows and nobody tells you before you pay for the generation, so the free in-page transpiler computes it. It reads a prompt into slots using the source model's grammar, re-emits it in the target's own labels and form, and reports the cost. What the target cannot render is taken OUT of the ported prompt rather than left in to be ignored, so the prompt and the loss report never contradict each other. Labels the source did not supply come out empty rather than invented, because a transpiler can only move what was there. It reports what the target does better too, since a move is not only a loss. And it catches the documented number-one failure when porting: quietly emitting one paragraph of prose for a model that requires labelled sections. Eleven model grammars, each with its required labels, its form, its word budget, its duration ceiling, its audio and dialogue capability, and the anti-pattern it invites. Five lanes: read it into slots, port it, compute what the move costs, verify the shape, or port to several models at once for a cross-model test. Lanes derived from the open-source MIT toolkit cclank/lanshu-awesome-ai-video-kit. Model capabilities change - check the vendor's own documentation before committing a budget. Not affiliated with or endorsed by cclank, ByteDance, Kuaishou, Alibaba, OpenAI, Google, MiniMax, Tencent, Pika or Runway.

Every public app is built from a security-scanned skill and must pass a clean scan — skill and frontend — before it can be listed. Have a skill of your own? Turn it into an app — or read the step-by-step walkthrough.