1:1 → 16:9
Portrait to cinematic landscape
Open the sides of a centered portrait for a website hero while continuing the room, window light, and depth of field.
Generative reframe for photos and product images
Uncrop & AI Expand Image
Use Uncrop & AI Expand Image to upload a photo, choose a wider or taller aspect ratio, guide the surrounding scene, and generate a new composition without cutting away the subject.
JPG, PNG, WebP 15 target frames 1K, 2K, 4K
Ready for your image
Expanded canvas
Ready for your image
Target 16:9
Preview example
Preview example Upload an image, then choose a target frame
The result creates a plausible new scene beyond the frame.
Original stays in this sessionWhat Does It Mean to Uncrop an Image?
To uncrop an image is to enlarge its canvas and generate visual context outside the current boundaries. The goal is a roomier or differently shaped composition without removing the subject to make it fit.
AI image expansion reads visible clues such as light, texture, perspective, depth, and repeated patterns, then proposes a plausible continuation around the source. Unlike stretching, it can add distinct details; unlike cropping, it can preserve the current subject while opening space around it.
This is generative reconstruction, not a recovery of the real scene. The model cannot know what was actually beyond the camera frame, so treat the output as a new creative asset and inspect it before use.
How to Uncrop & AI Expand Image
The focused workflow keeps the core decision simple: supply one source, choose the final shape, add only the scene direction that matters, then review the new composition.
- 01
Upload one source image
Choose a JPG, PNG, or WebP up to 20 MB. The source establishes the subject, crop, perspective, and lighting to continue.
- 02
Choose the target frame
Select a wider, taller, square, portrait, or panoramic aspect ratio that matches the final placement.
- 03
Guide the surrounding scene
Optionally describe what belongs beyond the original edges, such as a studio wall, sky, floor, foliage, or negative space.
- 04
Generate, inspect, and download
Pick 1K, 2K, or 4K, generate the expansion, compare it with the source, and download the result you want to keep.
Uncrop vs Crop vs Upscale
Uncrop, crop, and upscale solve different problems. Choose based on whether you need more scene area, less scene area, or more pixels inside the same scene boundaries.
| Decision | Uncrop / expand | Crop | Upscale |
|---|---|---|---|
| What changes | Adds newly generated context outside the visible frame | Removes outer content to tighten the frame | Increases pixel dimensions and may reconstruct detail |
| Best when | The image is too tight or needs a new aspect ratio | The frame contains distractions or excess space | The composition works but the file is too small |
| Composition | Becomes wider or taller without cutting away the subject | Becomes tighter and can remove parts of the scene | Usually keeps the same visible boundaries |
| New visual information | Yes—generated content appears beyond the original crop | No—existing content is removed | No new scene area; detail is enlarged or inferred inside the frame |
AI Image Expansion Examples
These examples cover three common reframing jobs: making a portrait wider, giving a product more vertical room, and turning a horizontal landscape into a story-shaped asset.
1:1 → 4:5
Product image to social portrait
Add height around a centered product for a 4:5 campaign placement while extending the set, shadows, and usable copy space.
16:9 → 9:16
Landscape scene to vertical story
Build new sky and foreground above and below a horizontal travel image to create a vertical story layout without slicing out the focal point.
Where AI Image Expansion Helps
Image expansion earns its place when the original content is useful but the current frame does not match the publishing surface. It is a composition tool before it is an effect.
Website heroes and banners
Turn a tight portrait or product crop into a wide banner with breathing room for a headline and call to action.
Social format adaptation
Reframe one approved asset for 1:1, 4:5, 9:16, or 16:9 placements without defaulting to a destructive crop.
Product campaign space
Extend a seamless set, tabletop, wall, or shadow area while keeping the product near the intended focal point.
Recovered composition options
Explore a less cramped frame around portraits, animals, architecture, and travel images when the original capture feels too close.
How to Get a More Natural Expansion
Natural-looking uncrop results depend on the evidence in the source, the amount of new canvas requested, and whether the direction agrees with the scene already visible at the edges.
Give the model enough evidence
Use a sharp source with visible texture, light direction, and environmental clues. A tiny or heavily compressed image leaves more of the continuation ambiguous.
Describe the edge, not a new story
Short guidance such as “continue the limestone wall and soft left window light” is easier to reconcile than several unrelated changes.
Expand in a plausible direction
Choose a ratio that creates space where the scene already implies it. Extreme reframing asks the model to invent more visual information.
Inspect every boundary
Check repeated patterns, architecture, shadows, limbs, text, logos, and the transition around the original frame before publishing.
What AI Uncrop Can and Cannot Do
The tool can generate context that feels visually continuous, but generative output remains a proposal. Knowing the boundary between expansion and factual recovery prevents the wrong expectations.
What it can do
- Generate new visual context around one source image.
- Adapt a composition to common landscape and portrait ratios.
- Use short scene guidance to influence the continuation.
- Produce higher-resolution output options for review and download.
What it cannot promise
- Recovery of the factual scene that existed outside the photo.
- Pixel-identical preservation of every original detail.
- Perfect text, logos, hands, reflections, geometry, or repeated patterns.
- Brush-selected, mask-based, or layer-based manual outpainting.
The current implementation uses the source, target ratio, resolution, and prompt to generate a new reframed image. If you need a broader whole-image change, use the AI Image Editor. If the frame is correct but the file needs more detail, compare the Topaz Image upscaler.
Banana Pro AI vs Pixelcut vs Adobe Firefly
Banana Pro AI, Pixelcut, and Adobe Firefly all support image expansion, but the surrounding workflow differs. The right choice depends on whether you want a direct task, custom canvas sizing, or an integrated editing suite.
| Decision | Banana Pro AI | Pixelcut Uncrop | Adobe Firefly Generative Expand |
|---|---|---|---|
| Primary workflow | Focused browser workbench: upload, choose a frame, add optional scene guidance, generate, compare, download | Browser uncrop flow with preset or custom canvas sizing | Generative Expand inside the broader Firefly editing workspace |
| Frame controls | 15 target ratios at 1K; compatible ratios at 2K and 4K | Common aspect-ratio presets plus custom size controls documented by Pixelcut | Freeform canvas resizing and aspect-ratio presets documented by Adobe |
| Creative direction | Optional plain-language guidance for the new surrounding scene | Prompt-assisted uncrop workflow with alternative generations | Optional text prompt, or leave it blank for an automatic continuation |
| Best fit | A direct uncrop task with visible credits, resolution choices, in-session source comparison, and download | A quick dedicated uncrop utility with custom canvas sizing | Teams already working inside Adobe’s connected editing workflow |
Choose Banana Pro AI
When you want a focused uncrop page, explicit resolution and credit controls, in-session source comparison, and direct download.
Choose Pixelcut
When a dedicated utility with custom canvas dimensions and multiple proposed continuations fits your workflow.
Choose Adobe Firefly
When Generative Expand should stay inside an Adobe canvas and continue into a broader creative editing workflow.
Uncrop and AI Expand Image FAQ
Straight answers about what uncrop means, how generated context differs from recovered pixels, supported files and ratios, credit use, and preservation limits.
What is AI uncrop?
AI uncrop expands the canvas beyond an image’s current boundaries and generates a plausible continuation around the source. It changes the composition without simply stretching existing pixels.
Is uncrop the opposite of crop?
In composition terms, yes: crop removes outer content, while uncrop adds newly generated visual context outside the visible frame. The added content is an AI interpretation, not lost pixels being restored.
Does AI image expansion recover what was outside the camera frame?
No. It creates a visually plausible continuation from clues in the source and your guidance. The result is not a recovery of the real scene that existed beyond the original photo.
Which files can I upload?
The workbench accepts JPG/JPEG, PNG, and WebP images up to 20 MB. Use the highest-quality source available when details and boundaries matter.
Which aspect ratios can I use?
At 1K you can choose 1:1, 4:5, 3:4, 2:3, 9:16, 4:3, 3:2, 16:9, 2:1, 21:9, 1:2, 3:1, 1:3, 9:21, or 5:4. At 2K and 4K, 4:5, 3:1, 1:3, 9:21, and 5:4 are unavailable.
How many credits does an expansion use?
The current workbench shows the cost before generation: 1K uses 20 credits, while 2K and 4K use 40 credits. Sign-in and available credits are checked when you generate.
Will the original subject stay pixel-identical?
Do not assume pixel-identical preservation. The workflow asks the model to preserve the subject and composition while generating a new output, so compare identity, products, logos, text, hands, geometry, and fine detail with the source.
Sources and Related Image Tools
Product facts are tied to the implemented workbench; competitor statements are limited to the vendors’ own published feature pages. Related tools cover whole-image editing and detail enhancement rather than repeating this page’s job.
Product comparison reviewed: 2026-08-24
Expand Your Image Beyond the Frame
Upload the best available source, choose the frame your layout needs, and add one clear clue about the surrounding scene.
Start expanding