OpenArt: Create consistent scenes using Omni Reference

OpenArt Try it →Watch full video →YYouri van Hofwegen

Demo summary

The narrator shows how to use Open Art's Omni Reference feature to maintain character consistency across different scene generations by referencing previous character and environment images.

Step-by-step

Paste your scene prompt into the prompt field
Click on the Omni Reference window
Go to the History tab
Select the character images to be used as references
Set all generation settings to the highest available options
Click Generate to create the first shot
For subsequent scenes, remove the previous reference image and add the most recently generated scene to provide context
Re-add the character images alongside the new scene reference and generate

Watch out for

You must select your character references for every scene they appear in to prevent them from morphing into different designs

Tips

Use the previous scene as a reference image so the AI has the full context of the environment and style you are working with
Include specific character moods in the prompt, such as 'serious' or 'happy', to make the scenes feel more alive

Highlights

“All the prompting is already taken care of. For the first scene, we're going to copy over the prompt and paste it in.”

All demos from “How to Make Animated Cartoon videos with AI (Full Course)”

2:523:02Generate consistent characters in Open ArtUsing the Nano Banana Pro model in Open Art, the narrator demonstrates generating character designs for an alien, a dog sidekick, and a slime monster by pasting detailed prompts and selecting aspect ratios.OpenArt· AI Person Generator
6:052:07Create consistent scenes using Omni ReferenceCurrentThe narrator shows how to use Open Art's Omni Reference feature to maintain character consistency across different scene generations by referencing previous character and environment images.OpenArt· AI Person Generator
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.OpenArt· Image to Video
10:201:06Smooth transitions with Grab Frame to VideoThe narrator demonstrates using the 'grab frame to video' feature to take the last frame of one clip and use it as the starting frame for the next in Kling 2.5 to ensure seamless animation transitions.KKling· AI Video Extender
12:081:21Create dynamic action sequencesThe narrator walks through building a fight scene by generating specific close-ups and action shots, such as a monster zoom-in and a character firing a laser, using image-to-video generation.LLuma AI· Image to Video
13:560:26Generate sound effects with Eleven LabsThe creator demonstrates the sound effects feature in Eleven Labs by prompting for a 'slimy monster punching' sound and playing back the generated results.ElevenLabs· AI Sound Effect Generator
Watch “How to Make Animated Cartoon videos with AI (Full Course)” →

AI Person Generator

45:571:49Build consistent AI charactersThe tutorial walks through the 'Build your character' workflow, selecting ethnicity, age, and clothing to create a persistent character named Oscar.Wade McMaster - Creator Impact
6:331:13Create a consistent AI character in OpenArtThe user builds a character named Amelia from a text prompt, generating a consistent reference set including front, closeup, and full-body views.Affiliation
7:580:31Generate new scenes with a saved AI characterThe user selects the previously created 'Amelia' character and generates a new image of her at a cafe, maintaining facial and hair consistency.Affiliation
1:210:42Maintain character consistency with Nano Banana 2The user demonstrates how to generate a scene with specific characters and then use the 'add visual references' feature to upload an image and maintain character consistency in a new prompt.Isa does AI
3:310:57Create consistent AI characters from image referencesThe user uploads an image of a red-haired woman to OpenArt to create a character reference, then applies that character to a new 'desert buggy' video prompt to maintain visual consistency.Eigi and AI
3:091:53Create consistent characters in OpenArtThe user demonstrates the 'Characters' feature in OpenArt, using text descriptions and the 'Pixar style' setting to generate consistent squirrel and fox characters.Roboverse
2:490:27Create a consistent character from text in OpenArtThe video shows how to use the 'Describe your character' feature and Nano Banana Pro model to generate and save a reusable character named Duncan.Roboverse
4:520:36Create a consistent AI characterThe video shows the process of creating and saving a consistent character in OpenArt using a text description and the Nano Banana Pro model.Isa does AI
4:540:34Generate a full-body character reference imageUsing the GPT Image 2 model in OpenArt, the creator generates a 16:9 reference image of the character in specific 'jungle combat gear' to lock in outfit and styling consistency.Isa does AI
5:310:51Generate consistent character angles with Flux Context MaxThe creator shows how to use the Flux Context Max model and image guidance to generate multiple consistent angles of the same character for training purposes.Roboverse
12:102:03Control character posture with 3D Pose EditorA detailed walkthrough of the 3D Pose Editor where the user manipulates a mannequin's limbs to set a specific pose for the AI influencer in a coffee shop.Roboverse
6:431:47Create consistent character sheets in OpenArtThe user shows how to use the character reference prompts from Grok in OpenArt with the Nano Banana 2 model to generate front and back views of characters like 'Lemonini' and 'Peachina' for visual consistency.Andrei Ai