demobook

OpenArt: Create video transitions with Start and End frames

Demo summary

The creator demonstrates how to use the 'Start Frame' and 'End Frame' feature in OpenArt with Google Veo 3.1 and Kling 2.5 to create a seamless transition between two different AI-generated scenes.

Step-by-step

  1. Generate or download an image to serve as your 'Start Frame'.
  2. Create and download a second image to serve as your 'End Frame'.
  3. Navigate to the Video section and select Image to Video.
  4. Select a model that supports start and end frames, such as Google Veo 3.1 or Kling 2.5.
  5. Upload your first image into the Start Frame slot.
  6. Upload your second image into the End Frame slot.
  7. Enter a simple prompt describing the transition between the two scenes.
  8. Click generate to create the video transition.

Options

  • Use Google Veo 3.1
  • Use Kling 2.5
  • Use Luma Dream Machine (1 2.5)

Watch out for

  • Not every video model supports the Start and End frame feature (e.g., Sora 2 does not have it).
  • Luma 1.2.5 may not always end precisely on the provided end frame.

Tips

  • Keep your transition prompts simple as it often works better for video generation.
  • If the first result isn't perfect, try regenerating with the same settings to get a better result.
  • Ensure your aspect ratio (e.g., widescreen) matches across your generated images and video settings.

Highlights

that transition is sick

All demos from “How to Start Making AI Videos in 2026

  1. 5:060:57Generate AI image from PDF prompt in OpenArtThe creator demonstrates how to upload a PDF to ChatGPT to extract a prompt and then pastes it into OpenArt using the Nano Banana Pro model to generate a 4K image of a person on the moon.OpenArtText to Image
  2. 8:291:01Image-to-image generation with Seedream 4.5The creator uses a personal photo as a reference in OpenArt to generate a cinematic movie character floating in a bubble using the Seedream 4.5 model.OpenArtImage to Image
  3. 15:322:48Create video transitions with Start and End framesCurrentThe creator demonstrates how to use the 'Start Frame' and 'End Frame' feature in OpenArt with Google Veo 3.1 and Kling 2.5 to create a seamless transition between two different AI-generated scenes.OpenArtImage to Video
  4. Watch “How to Start Making AI Videos in 2026” →

Image to Video

  1. 9:300:40Animate scenes with Kling 2.6The creator demonstrates animating a static image into a 5-second video clip of characters walking using the Kling 2.6 model within the Open Art interface.Youri van Hofwegen
  2. 2:300:40Animate images using Kling 2.6 in OpenArtThe demonstration shows uploading a static image and using the Kling 2.6 model to generate a realistic video of a girl drinking coffee with natural steam and background movement.Affiliation
  3. 3:030:43Image-to-video conversion using Kling 2.6A demonstration of uploading a static photograph of a city street and using the Kling 2.6 model to animate it with rain and motion effects.Thomas Creates
  4. 4:400:42Create fight scene from image with Wan 2.6The user uploads a Tekken 8 reference image and uses Wan 2.6 to animate a cinematic fight scene between two characters.Ai Chemy
  5. 5:450:50Image-to-Video workflow in OpenArt to save creditsThe user demonstrates generating a high-quality image of a gold bar first using the 'Image Gen 4' model, then uploading that image to the VEO3 video generator to ensure visual consistency before animating the slice.unknown
  6. 3:450:44Image-to-video with start frame in Veo 3.1The user uploads a starting frame image and a prompt to guide the Veo 3.1 model in generating a consistent 4K video sequence.Youri van Hofwegen
  7. 4:420:23Image-to-video with start frame in LTX 2.3The LTX 2.3 model is shown processing a starting frame image to generate a high-resolution video with slow-motion effects.Youri van Hofwegen
  8. 6:430:30Multi-image reference video generation with Veo 3.1The demonstration shows how to use a character sheet and an object reference image simultaneously in Veo 3.1 to maintain visual consistency.Youri van Hofwegen
  9. 8:090:45Reference image workflow with SeaDance 2.0The SeaDance 2.0 model is shown generating a 720p video using a character sheet, boxing glove reference, and a starting frame.Youri van Hofwegen
  10. 31:141:10Sync motion from one video to anotherThe creator uses Motion Sync to apply the movements of a boxing video to a static cyberpunk character image.Wade McMaster - Creator Impact
  11. 6:020:33Image-to-video generation with Vidu 3.1The user demonstrates how to use a previously generated image as a foundation for video, using the Vidu 3.1 model to animate specific motions.Roboverse
  12. 6:580:43Animate images using Cling 3The creator demonstrates converting a still image into a 5-second video using the Cling 3 model within OpenArt's video interface.Isa does AI