How to Extend an Image with AI Outpainting
To extend an image with AI, upload it to bgeraser's generative expand tool, which uses outpainting to seamlessly generate new content beyond the original borders of your photo. Need more sky above a landscape, more floor beneath a product, or a wider aspect ratio for a social media banner? The AI analyzes the existing content and generates matching scenery, textures, and lighting in the extended areas. The result looks like the photo was originally captured with a wider frame.
Step-by-Step Instructions
Upload your image
Drop your photo into the upload area. The generative expand tool works with JPG, PNG, and WebP files. This tool is most commonly used when you need to change an image's aspect ratio without cropping, when a photo is framed too tightly around the subject, or when you want to add more context to a scene. Portrait photos that need to become landscape banners are a particularly common use case.
Choose which direction to extend
Select which edges of the image you want to extend: top, bottom, left, right, or any combination. You can specify how many pixels to add in each direction or choose a target aspect ratio and let the tool calculate the expansion automatically. For example, converting a 4:5 portrait to a 16:9 banner would extend the left and right sides significantly.
AI generates the new content
The AI analyzes the existing image for context: lighting direction, color palette, textures, perspective, and scene content. It then generates photorealistic content in the extended areas that seamlessly blends with the original. For a landscape, it continues the sky gradient, terrain features, and atmospheric haze. For an interior, it extends walls, floors, and ambient lighting. Processing takes 10 to 30 seconds depending on how much new area is being generated.
Review, refine, and download
Inspect the result, paying attention to the transition zone between original and generated content. If the blending is not seamless in an area, you can regenerate for a different result since the AI produces a different variation each time. Once satisfied, download the extended image at full resolution. The original portion of the image remains pixel-perfect; only the extended areas contain generated content.
Tips for Best Results
Generative expand works best when extending natural scenes (sky, water, grass, interiors) and worst when extending faces or text. If your subject is near the edge, the AI may attempt to extend the person, which produces uncanny results. Crop so the subject has some margin before extending.
For social media repurposing, keep a set of target aspect ratios bookmarked: 1:1 for Instagram posts, 4:5 for Instagram portraits, 16:9 for YouTube thumbnails, and 9:16 for Stories and TikTok. Generative expand lets a single photo hit all four ratios.
Use a text prompt to guide the generated content if the default result is not what you want. Describing the desired scenery (such as 'clear blue sky' or 'wooden floor') gives the model direction for what to fill in the extended area.
Frequently Asked Questions
What is AI outpainting?
Outpainting is the process of extending an image beyond its original borders by generating new content with AI. Unlike cropping (which removes content) or scaling (which stretches existing content), outpainting creates new pixels that are contextually appropriate for the scene. The term comes from the opposite of 'inpainting,' which fills in regions inside an image.
Can I extend a portrait photo to landscape orientation?
Yes. This is one of the most popular uses of generative expand. Upload a vertical portrait, and the tool generates natural-looking content on the left and right sides to create a horizontal landscape version. The AI matches the background environment, lighting, and color tone of the original photo so the extension looks seamless.
Does the generated content look realistic?
For natural environments, interiors, and product backgrounds, the generated content is highly realistic and often indistinguishable from the original photo. Results are best for areas with organic textures like sky, foliage, water, and walls. Structured elements like architecture, signage, or people may occasionally show artifacts.
How much can I extend an image?
You can extend an image up to 50% of its original dimensions in any direction. For a 1000x1000 image, that means adding up to 500px on each side. For larger extensions, you can run the tool multiple times, extending the result of the previous pass. Each pass maintains quality because the AI generates coherent content, not stretched pixels.
Will the extended areas match the original image quality?
The generated content matches the visual quality and style of the original image. If your input is a high-resolution photo, the extended areas will have corresponding sharpness and detail. If the input is low quality, the extensions will match that level. The AI does not upscale or enhance the original portion; it focuses on making the new content blend seamlessly.
Ready to Try It?
Drop an image below to get started. No signup required for your first images. 15 free credits every month.