Social Media Image Resizer
Resize photos in your browser for Instagram feed, Stories, Reels, TikTok, and YouTube thumbnails. Files stay on your device.
100% private - images are resized in your browser and never uploaded.
1080 × 1080 px
Fills the frame and crops overflow
Runs in your browser. Nothing you enter is uploaded.
Overview
Cropping picks a shape; resizing produces an exact pixel size. When a platform documents 1080×1350 or 1584×396, guessing at a ratio is not enough - this tool renders your image onto a canvas of precisely those dimensions and exports it.
The presets cover the sizes people look up most: Instagram square, portrait, and Story, TikTok video, YouTube thumbnail, X post, Facebook cover, and LinkedIn banner. Custom accepts any width and height up to 8000 pixels. Processing is local - the image is decoded, drawn, and exported inside the browser with no upload.
Cover versus contain, and when the background colour matters
Cover scales the image until it fills the target frame completely and crops whatever hangs over the edges. It is the right choice for photographs, where a full-bleed result matters more than seeing every pixel of the original. Contain scales the image until the whole thing fits inside the frame, leaving empty space on two sides which is filled with the background colour you choose.
Contain earns its place with logos, product shots on white, and any graphic where cropping would cut something meaningful. A 1:1 logo forced into a 1584×396 LinkedIn banner via cover would be reduced to a horizontal slice of itself; via contain it sits intact and centred on a chosen colour. Matching that colour to the brand background rather than leaving it white is what makes the letterboxing read as intentional.
Zoom and the two pan sliders apply in both modes, so a cover crop can be repositioned before export and a contained image can be nudged off centre. Reset position returns all three to their defaults in one click.
How the output canvas is produced
Selecting a preset sets the target width and height directly; Custom reads your two number fields, clamped to at least 1 pixel. An offscreen canvas is created at exactly those dimensions, filled with the background colour, and the decoded image is drawn onto it using the selected fit mode with your zoom and pan applied to the draw geometry.
The visible preview is that same canvas redrawn at up to 480 pixels with high-quality image smoothing enabled, so the proportions and framing you see are exactly what exports. The download writes the full-size canvas - not the preview - as a PNG named with the target dimensions.
Because the canvas is created at the target size and the source is drawn into it in a single operation, there is no intermediate resampling step to soften the result. Enlarging a small source beyond its native resolution still interpolates, though: the tool will happily draw a 400-pixel image into a 1920-pixel frame, and it will look like a 400-pixel image did.
A square logo placed on a LinkedIn banner
Input
Source: 1000 × 1000 logo on transparent background Preset: LinkedIn Banner (1584 × 396) Fit mode: contain · Background: #0B1120 · Zoom 1.00×
Output
Original 1000×1000 · Output 1584×396 · Mode contain · Zoom 1.00×
Cover mode here would have scaled the logo to 1584 pixels wide and then kept only the middle 396-pixel horizontal band - a slice through the centre of the mark. Contain scales it to 396 pixels tall, centres it, and fills the remaining width with the dark brand colour, producing a banner that looks designed rather than cropped.
Limitations
- Export is PNG only. Photographic output will be large compared with JPEG and usually warrants a pass through the compressor.
- Enlarging past the source resolution interpolates rather than adds detail, so upscaling a small image produces a soft result at any target size.
- Aspect ratio is not locked in Custom mode. Entering mismatched width and height simply crops harder in cover mode; the image is never stretched.
- No batch processing - one image at a time, with settings that do not carry over between files.
- The background colour applies only in contain mode; in cover mode it is never visible because the frame is fully covered.
- Preset dimensions reflect commonly published platform recommendations, which the platforms change without notice.
When not to use this tool
- If you only care about the shape and want the tool to pick sensible output dimensions, the image cropper is quicker - ratio presets, no width and height fields.
- If the image is already the right size and only the file is too big, compress it instead of re-rendering it.
How to use it
- 1
Choose the preset for your destination, or select Custom and enter exact width and height.
- 2
Pick cover for photographs and contain for logos or graphics, setting the background colour to match your brand when using contain.
- 3
Use the pan sliders to position the subject inside the frame, and zoom only when the composition needs it.
- 4
Download the PNG at the target dimensions, then compress it if you are uploading somewhere with a file size limit.
Questions
Which fit mode should I use for a photograph?
Cover, almost always. A photograph reads better filling the frame than floating on a coloured band, and the pan sliders let you choose which part survives the crop. Reserve contain for logos, diagrams, and product cut-outs where losing an edge would lose meaning.
Will resizing to a larger preset improve quality?
No. Drawing a 600-pixel image into a 1920-pixel canvas produces a 1920-pixel file that carries no more detail than the original. Start from the largest source you have; if the only copy is small, a smaller target size will look sharper than an enlarged one.
Why does my output look different from the platform preview?
Platforms apply their own crops and overlays on top of what you upload - Stories place UI at both ends, LinkedIn banners are cropped differently on mobile than on desktop. Export at the documented size, then check it in the app before committing.