Reframe
Intelligently reframe and crop images using AI — change aspect ratios, zoom out, or shift composition while the AI seamlessly fills in what lies beyond the original frame.
Overview
Reframe is an AI-powered composition tool that lets you change the framing of any image — expanding the canvas, shifting the focal point, or converting between aspect ratios — while the AI generates contextually accurate content to fill the newly added areas. The result looks as if the photo was always composed that way.
Key Benefits
- Aspect ratio conversion: Switch between portrait, landscape, square, and cinematic formats in seconds
- Smart zoom-out: Reveal more scene around the subject without distorting the original content
- Contextual AI fill: The model imagines what lies beyond the original frame in a visually consistent style
- Non-destructive workflow: The original image remains untouched; only the canvas extension is generated
Intelligent Composition
AI understands the scene's lighting, perspective, and style before generating any new content, ensuring seamless blending.
Contextual Extension
The model imagines plausible content beyond the original edges — sky, ground, walls, foliage — that matches the existing image.
Aspect Ratio Conversion
Convert from any ratio to any other ratio while keeping your subject centered and the composition balanced.
Smart Zoom Out
Expand the visible scene proportionally in any direction while preserving the original pixel content at full quality.
Focus Shifting
Move the subject's position within the frame — apply rule-of-thirds, centre the focal point, or offset for creative effect.
Seamless Blending
Generated pixels blend imperceptibly with the original — no visible seams, colour mismatches, or perspective distortion.
How to Use Reframe
Reframing an image takes four simple steps:
Step 1 — Upload your image
Drag and drop, click to browse, or paste an image URL. Supported formats: JPG, PNG, WebP. Maximum file size: 10 MB.
Step 2 — Choose your reframe type
Select from Zoom Out, Aspect Ratio Change, or Focus Shift. Configure the target dimensions or ratio using the controls. You can preview the crop boundary on the original image before the AI generates any fill content.
Step 3 — AI processing
The AI analyses the image's scene, lighting, colour palette, and perspective. It then generates new content for the extended canvas areas, blending the boundaries imperceptibly. Processing typically takes 30–60 seconds.
Step 4 — Download the result
Preview the reframed image at full resolution. If unsatisfied, regenerate with a different configuration. When ready, download in your chosen format.
Pro Tip
For the most natural zoom-out results, keep the extension to 1.5× the original size or less. Larger extensions require the AI to invent more content, which increases the chance of visual inconsistencies at the edges.
Reframe Types & Aspect Ratios
Reframe supports three distinct composition operations, each suited to different use cases:
Zoom Out
Expands the visible scene uniformly around the subject. Ideal for restoring context that was cropped from an original photo, or for creating a wider establishing shot from a tight frame.
Aspect Ratio Change
Converts the image to a target ratio by extending the shorter dimension. The AI fills the added strips on either side (or top/bottom) with contextually appropriate content.
| Format | Ratio | Ideal for |
|---|---|---|
| Square | 1:1 | Instagram posts, profile images, thumbnails |
| Portrait | 3:4 / 4:5 | Mobile content, Instagram Stories, Pinterest |
| Landscape | 16:9 | Web banners, YouTube thumbnails, presentations |
| Cinematic | 21:9 | Wide displays, film stills, hero images |
| Custom | Any | Print layouts, advertising specifications |
Focus Shifting
Repositions the subject within the frame without resizing the canvas. The vacated areas on the opposite side are filled by the AI. Useful for applying rule-of-thirds composition to an originally centred photo.
Best Practices
Source Image Tips
- Use high-quality originals: The AI cannot invent detail that is absent in the source; start with the highest-resolution image available
- Prefer well-lit images: Consistent, even lighting makes the AI extension blend more naturally at the boundaries
- Avoid heavy compression: JPEG artefacts are replicated and amplified in the generated areas; use PNG or a high-quality JPEG
- Scene context helps: Images with recognisable environments (outdoor landscapes, interiors, streets) produce more coherent extensions than abstract or studio-only shots
Configuration Tips
- Keep extensions moderate: Extensions beyond 2× the original dimension increase generation time and reduce consistency
- Use Regenerate: If the first result has inconsistencies, regenerate — the model introduces randomness and a second pass often resolves edge issues
- Combine with Remove Background: Remove the background first, then reframe onto a new backdrop for full compositional control
Known Limitations
- Very abstract or pattern-based images (e.g. solid textures, geometric designs) may produce repetitive or inconsistent extensions
- Images with text near the edges may have the text cut off or incorrectly extended — manually crop text out before reframing
- Processing time increases proportionally with the amount of new canvas area being generated