Text-to-video

• A great way to start is to write short and simple prompts, such as:

  • underwater
  • fireworks above the city
  • a ladybug climbing on a leaf

• Adding styles to the beginning of the prompt can add a specific aesthetic to the shot: for example: cinematic, close-up shot of an old pirate captain at sea

  cinematic_close-up_shot_of_an_old_pirate_captain_at_sea.mp4

  anime style, a ginger cat, beautiful sky sunset

  anime style a ginger cat beautiful sky sunset.mp4

  3d render, a pigeon in a subarban neighorhood

  3d render a pigeon in a subarban neighorhood.mp4

• If you want to be more advanced, add simple descriptive words in the prompt: define the camera motion, the surroundings, etc. A good prompt structure to follow is:
  *style + camera motion + (description of) Main character/ subject + verb.+ surroundings + additional info*

• For example:

• Simple is key. Some models perform well with detailed complex prompts. The Pika model performs better with simple prompts.

• For example:

  • Anime, a girl reading under the sakura tree
  • A koala as DJ performing at a music festival
  • Aerial drone shot of a traditional Japanese temple covered in snow

• Good style keywords to use:

• Camera Keywords that work well:

• Text-to-Video examples:

Image to Video

We highly recommend image to video as it can produce incredible results. adjusting the strength of motion parameter and adding the camera motion parameter can change the amount of motion drastically. Text prompt is optional here.

• Upload an image that you want to animate and add a prompt
• There are different ways to animate the image, you can try the following approaches:

  1. describe the specific subject you want to animate in an image. for example, Input Image:

     

     Prompt: owl

     owl_seed6445027359260674.mp4

  2. describing the action you want to animate in an image Prompt: owl looking at the camera

     owl_looking_at_the_camera_seed408550403213911.mp4

  3. If you are using an AI generated image, try using the original image prompt along with added camera motion and strength of motion parameters. for example, prompt that was used to generate the image: a red owl sitting on a branch in a cyan and pink forest, cinematic shot, shot with red cameras and prime lens

     added parameters: "cameraParameters": {"zoom": "in"}, "cameraMotionStrength": 4

     a_red_owl_sitting_on_a_branch_in_a_cyan_and_pink_forest,_cinematic_shot,_shot_with_red_cameras_and_p_seed2601356747359434.mp4

  4. Describe the action and camera motion you’re expecting in the prompt along with a brief description of the scene and add a camera motion/strength of motion parameters.

     • i.e: Input Image:

       

       Prompt: slow dolly in, perfume bottle in a forest, flare
       Parameters: "cameraParameters": {"zoom": "in"}, "cameraMotionStrength": 2

       slow_dolly_in,_perfume_bottle_in_a_forest,_flare_seed8555752376115185.mp4

       Prompt: drone shot of a winery, circling from above
       Parameters: "cameraParameters": {"rotate": "cw"}, "cameraMotionStrength": 3

       drone_shot_of_a_winery,_circling_from_above_seed8570214566508136.mp4

  5. Let our AI do the work, just upload the image, no prompt needed for example: No Prompt, only Parameters: "cameraParameters": {"zoom": "in"}, "cameraMotionStrength": 2

     _seed2430350719456520_upscaled.mp4

Video to Video:

• Upload a video that you want to change to another style