Skip to main content

Prompt guide · Updated October 17, 2026

The Crazy Frog AI Prompt for Every Video Tool

The Crazy Frog AI prompt works together with a reference video: it tells the model to copy the trend clip’s camera, cop and flip, and to put the person from your photo on the invisible motorcycle. One line of text is not enough. Below is the exact prompt behind this site, with versions for Kling and Veo.

Copy and paste

Crazy Frog AI prompts for 3 video tools

Pick your tool, pick a scene, copy. The Seedance prompt is word for word what this site runs; the video above came from it with the Snowy pass line.

  1. 1Open a Seedance 2 tool that takes a reference video and a reference image
  2. 2Add a trend clip of 15 seconds or less as the reference video and your photo as the reference image
  3. 3Paste the prompt, set 9:16 and 15 seconds, and generate
Skip the prompt, upload a photo

Crazy Frog AI runs the Seedance prompt with its own reference ride. You only add a photo.

Recreate the reference video exactly: same camera movement, framing, timing, the police officer on the white hover-bike, the jump and the landing. Replace only the rider in the grey hoodie with the person from the reference image: use that person's face, hair, skin tone and outfit. The rider keeps the low crouched riding pose with both fists out, and there is still no vehicle under the rider. Do not change the police officer. Change the location to a winding snowy mountain pass road at blue hour, deep snowbanks, pine trees and light falling snow. No text, no watermark.

Our reference ride has a rider in a grey hoodie. If your clip differs, change "the rider in the grey hoodie" to what that rider wears, so the model knows who to replace.

Line by line

What each part of the Crazy Frog AI prompt does

Every sentence fixes one thing a video model gets wrong on its own. Hover or tap a note to light up its words.

Recreate the reference video exactly: same camera movement, framing, timing, the police officer on the white hover-bike, the jump and the landing. Replace only the rider in the grey hoodie with the person from the reference image: use that person's face, hair, skin tone and outfit. The rider keeps the low crouched riding pose with both fists out, and there is still no vehicle under the rider. Do not change the police officer. {scene} No text, no watermark.

Tested

Why a one-line Crazy Frog AI prompt fails

Most prompt posts boil the trend down to one sentence. We ran exactly that so you do not have to.

One sentence, one photo, no reference video

We gave Seedance 2 Mini the same photo as the video at the top and one sentence: "Crazy Frog AI trend video of this person riding an invisible motorcycle down a highway, chased by a cop." This is what came back.

  • A see-through glass motorcycle: the model takes "invisible" literally
  • A police car instead of the hover-bike cop
  • No low crouch and no flip, just a seated ride
  • A close camera that never shows the chase

Tips

Tips for a Crazy Frog AI prompt that works

  • Say there is no vehicle

    The word "motorcycle" pulls one into the frame. "There is still no vehicle under the rider" keeps it out.

  • Name the person to replace

    Describe the original rider by their clothes, such as "the rider in the grey hoodie", so the face lands on the rider and not the cop.

  • Match the reference length

    Ask for the same duration as your reference clip, 11 to 15 seconds, so the flip is not cut off.

  • Change one thing at a time

    Swap only the scene sentence between takes. Rewriting the whole prompt changes the motion too.

  • Set 9:16 from the start

    Generate vertical. Cropping a horizontal video later cuts the cop out of the frame.

  • Add the sound last

    Generate without music and add the Crazy Frog Axel F track in TikTok, so your clip shows up on the trend’s sound page.

FAQ

Crazy Frog AI prompt FAQ

Skip the prompt

Crazy Frog AI runs this prompt with its own reference ride. Upload a photo and get the video in 3 to 6 minutes.