bria/expand-image

Bria Expand expands images beyond their borders in high quality. Resizing the image by generating new pixels to expand to the desired aspect ratio. Trained exclusively on licensed data for safe and risk-free commercial use

368.9K runs

Readme

Expand images beyond their original frame with high visual fidelity.

BRIA Expand Image is an image outpainting model that generates new pixels around an existing image to fit a larger canvas or a new aspect ratio. It is designed for commercial workflows and built on licensed data.

Overview

Use BRIA Expand Image to turn tightly cropped images into flexible, format-ready assets. Instead of cropping important content, the model extends the scene naturally around the original image so you can adapt one asset into multiple layouts.

It is well suited for marketing creatives, ecommerce imagery, social media formats, banners, hero images, and any workflow that needs extra space for composition, text, or channel-specific sizing.

Key Benefits Expand images instead of cropping them Adapt assets to new aspect ratios such as 1:1, 4:5, 16:9, and 9:16 Guide the generated areas with a text prompt Preserve consistency with deterministic seed control Support commercial usage with a model trained on licensed data How It Works

BRIA Expand Image keeps the original image as the anchor and generates new surrounding content to match a desired output format. You can either:

expand to a target aspect ratio, or define a custom canvas size and control exactly where the original image is placed

An optional prompt can be used to guide the newly generated regions. If no prompt is provided, the system can automatically infer one from the image.

Ideal Use Cases Reframing product photos for ads and marketplaces Turning landscape images into vertical assets for stories and reels Expanding lifestyle or branded imagery for banners and landing pages Creating extra negative space for copy, UI, or overlays Producing multiple aspect-ratio variants from a single source image

Model created
Model updated