Skip to content

Crop image

Cut a photo to the part and shape you want.

Drop an image here

JPG, PNG, WebP, AVIF, HEIC, SVG and GIF

How to use Crop image

  1. Add an image. Drop a JPG, PNG, WebP, AVIF, HEIC, or SVG file into the drop zone.
  2. Choose a Shape preset. Click Free or lock the frame to 1:1, 4:3, 3:2, 16:9, 9:16, or 4:5.
  3. Frame your subject. Drag the box or corner handles, or tab to the crop area to nudge with Arrow keys.
  4. Crop and save. Click Crop image to download the trimmed file in its original format.

Your files are processed on this device and never uploaded. They stay private, and there's nothing to delete afterwards.

Choosing the right aspect ratio from Free to 16:9 and 4:5

Trimming away distracting edges improves composition and prepares photos for strict platform templates without stretching subjects. Under Shape, seven buttons switch between Free form selection and six locked aspect ratios tailored to photography, video, and social feeds.

Locked social presets (1:1, 4:5, 9:16, 16:9) and print ratios

Clicking 1:1, 4:3, 3:2, 16:9, 9:16, or 4:5 locks corner handles to that exact mathematical proportion while keeping the box centered inside the image bounds. Even if you drag a single corner diagonally, the opposite axis scales proportionally to preserve the target ratio.

Social platforms crop mismatched uploads automatically, often slicing off heads or headlines. If you need circular profile badges instead of rectangular frames, you can create round avatar cutouts after trimming the background.

Freeform trimming with the 3x3 composition grid

Selecting Free unlocks independent width and height adjustments across all four corner handles (nw, ne, sw, se). A subtle 3×3 grid overlays the active selection so you can align horizons and eyes along rule-of-thirds intersections.

Shape preset Aspect ratio Typical publishing use case Recommended pixel target
Free Unconstrained Custom document scans and UI screenshots Native source pixels
1:1 Square (1.00) Profile avatars and square catalog cards 1,080 × 1,080 px
4:5 Portrait (0.80) Tall Instagram and LinkedIn feed posts 1,080 × 1,350 px
9:16 Vertical (0.56) Mobile Stories, Reels, and Shorts covers 1,080 × 1,920 px
16:9 Widescreen (1.78) YouTube thumbnails and presentation slides 1,920 × 1,080 px
3:2 / 4:3 Classic camera DSLR 4×6 prints and tablet slideshows 1,600 × 1,200 px

How pixel slicing and keyboard precision controls work

Mouse and touch dragging cover quick framing, while keyboard shortcuts let you align selections down to a few pixels without fighting pointer jitter. Every movement updates the New size counter immediately below the Shape buttons.

Moving and resizing the frame with Arrow keys and Alt modifiers

Focus the crop box with Tab and press any Arrow key to shift the frame by 2% (0.02) of the image dimension. Holding Alt while pressing an Arrow key reduces each step to 0.5% (0.005) for tight alignment around borders or text.

To resize from the center using the keyboard, press Shift plus Arrow keys when Free is active, or tap + and - to expand and shrink locked ratios symmetrically.

Live New size pixel readout and 3% minimum selection bounds

The New size pill calculates Math.round(box.w * imgW) × Math.round(box.h * imgH) px continuously as you drag. Enforcing a 0.03 (3%) minimum box fraction ensures corner handles never invert or collapse onto each other.

As documented in the MDN CanvasRenderingContext2D.drawImage() documentation, the tool passes the exact (sx, sy, sw, sh) source rectangle into drawImage, extracting pixels directly in your browser. If you need to divide a wide panorama into sequential carousel slides instead of keeping one crop, you can slice one photo into Instagram grid tiles.

Does cropping an image reduce resolution or sharpness?

Cropping removes outer pixels rather than shrinking the subject inside your selection box, so the remaining area retains 1:1 native pixel detail. Only selections exceeding 4,096 px on a side or 16,777,216 px total are scaled down to fit browser canvas limits.

1:1 source pixel extraction versus canvas dimension caps

When you trim a 4,000 × 3,000 px camera photo down to a 1,800 × 1,800 px square, the output canvas allocates exactly 1,800 × 1,800 px with zero downsampling. Per the MDN imageSmoothingQuality documentation, setting imageSmoothingQuality to "high" keeps edges clean whenever clampCanvasSize scales down massive panoramas above 4,096 px.

Tip: Cropping changes pixel dimensions by cutting away margins, not by scaling. If you need an exact target width like 800 px after framing your subject, pass the output straight to Resize image using the next-step button.

Format preservation, 0.92 encoding quality, and -cropped filenames

The exporter calls sameFormat so a .png stays .png, a .webp stays .webp, and a .jpg stays .jpg, appending -cropped before the extension. Lossy exports encode at 0.92 quality via encodeWithFallback, balancing clean gradients against compact file sizes.

If a browser cannot encode an exotic input format such as AVIF or HEIC directly, encodeWithFallback steps down cleanly to WebP or PNG instead of failing. Your original file on disk remains completely untouched while the browser saves a fresh -cropped copy.

EXIF orientation normalization and transparent canvas fills

Phone cameras frequently store portrait shots sideways in raw sensor buffers and attach an EXIF orientation tag. Cropping raw unrotated buffers would place your selection on the wrong edge unless orientation is resolved first.

Upright portrait decoding before setting crop coordinates

According to the MDN createImageBitmap() documentation, calling createImageBitmap with imageOrientation: "from-image" bakes EXIF rotation into the decoded bitmap. Both the preview frame and the final pixel slice share identical upright coordinates.

If a scanned document or landscape shot still needs a 90-degree turn or horizontal flip before framing, open the image rotation and mirroring tool first.

Keeping PNG transparency versus matte fills for JPG exports

When cropping a transparent PNG or WebP sticker, needsBackground returns undefined, leaving the alpha channel intact around the trimmed subject. For opaque formats such as JPG, the canvas paints a #ffffff backdrop before drawing so transparent edges never turn solid black.

Once you finish cropping, the result panel offers one-click Resize and Compress handoff buttons that carry your trimmed image directly into the next tool without re-picking files.

Questions

Which aspect ratio should I use for social media posts and banners?

Use 1:1 for square feed posts and profile avatars, 4:5 for tall portrait feed photos, 9:16 for vertical Stories and Reels, and 16:9 for widescreen video thumbnails and header banners.

Does cropping a photo reduce its image quality?

The cropped region keeps its native 1:1 pixel resolution as long as the trimmed width and height stay within the 4,096-pixel canvas cap. Lossy formats like JPG and WebP re-encode at a high 0.92 quality setting to prevent visible blockiness.

How do I crop an image precisely using keyboard arrow keys?

Press Tab to focus the crop box, then use Arrow keys to move in 2% steps or hold Alt with Arrow keys for 0.5% micro-nudges. Hold Shift with Arrow keys to adjust width and height, or press + and - to scale from the center.

What file format does the cropped image export as?

The tool inspects the input file via sameFormat and exports in the matching format whenever the browser supports encoding it, appending -cropped to the base filename. Transparent PNGs stay PNGs with intact alpha channels.

Why is the minimum crop selection limited to 3% of the image?

Enforcing a 3% minimum box fraction prevents accidental zero-pixel drags or collapsed handles when resizing corners quickly on touchscreens or trackpads. Even on a 1,000-pixel photo, you can still isolate a 30-pixel region.

Can I crop transparent PNG graphics without losing transparency?

Yes, cropping a PNG keeps the full alpha transparency channel untouched. Only opaque formats like JPG fill the canvas with a solid white background before saving.

Updated