Scenario

Scenario allows you to train custom LoRA models using your own image datasets. You upload 15-50 reference images, and the platform trains Flux Dev LoRA or SDXL LoRA models that capture your specific style, characters, or brand assets. The training process typically takes 20-40 minutes and results in a model that generates consistent, on-brand outputs aligned with your creative vision.

Yes, Scenario provides a comprehensive RESTful API that enables programmatic control over model training, asset generation, and content management. You can integrate it directly into Unity, Unreal Engine, or custom pipelines. The API supports batch processing, webhook notifications, and OAuth2 authentication for secure integration with your existing development workflows.

Compute Units (CUs) are the billing metric for priority processing—for example, generating a Flux image costs about 5 CUs. Relax Mode, available on Pro and higher plans, allows unlimited generations without consuming CUs, but requests are processed when server capacity is available. This makes Relax Mode ideal for exploratory work while saving CUs for time-sensitive production assets.

Scenario uses custom-trained LoRA models and advanced prompt management to ensure visual consistency. By training on your specific art style or brand assets, the platform learns the exact characteristics, color palettes, and aesthetic elements. Combined with reference image conditioning and the Activity Dashboard for tracking generation parameters, teams can reproduce identical styles across unlimited asset variations.

Scenario provides access to leading video models including Google Veo, Runway Gen-3, and Luma Dream Machine, plus 25+ 3D generation models for creating meshes, textures, and materials. The platform supports text-to-video, image-to-video, video editing, upscaling to 4K, and 3D model generation from images. All models are accessible through the same interface with consistent workflows.