Image Upscaler
2x · 3x · 4x Browser-OnlyEnlarge images without quality loss using Lanczos resampling — runs entirely in your browser, no uploads, no AI APIs.
Image Input
Drop image here Release to upload
or click to browse
JPG
PNG
WebP
Max 25 MB
Upscale Factor
Output will be large — processing may take a few seconds on slower devices.
Resampling Algorithm
Lanczos produces the sharpest results; Bilinear is fastest.
Output Format & Quality
Upload an image to get started
Drag & drop or click the upload area on the left to select a JPG, PNG, or WebP image.
All processing happens locally in your browser — your image is never uploaded.
This may take a moment for large images
Upscaled Result
Before / After Comparison
Drag the divider
Select your upscale factor and algorithm on the left, then press Upscale Image.
🎨
Lanczos
Sharpest result
🔒
Private
Stays in browser
⚡
Fast
Web Workers
📥
PNG / JPEG
Export formats
Category
🖼️ Image ToolsHow It Works
- Upload your JPG, PNG, or WebP image
- Choose the upscale factor (2x, 3x, or 4x)
- Select a resampling algorithm
- Click Upscale Image
- Compare before/after, then download
Related Tools
About This Tool
The Image Upscaler enlarges pictures by a factor of two, three or four using classical resampling algorithms implemented in JavaScript and run entirely in your browser. Three algorithms are offered: Lanczos-3, which gives the sharpest result and is the default, bicubic using the Keys formulation, and bilinear, which is the fastest. You can preview the output dimensions for each scale factor before committing, and once processing finishes a compare view lets you switch between the original and the upscaled version to judge the difference. Output can be saved as lossless PNG, or as JPEG or WebP with an adjustable quality setting. The panel reports the source dimensions, the resulting dimensions, the output file size and how long the upscale took in milliseconds. No AI model and no external service is involved, so nothing is uploaded and the same result is produced every time.
Frequently Asked Questions
No. It applies classical resampling filters, Lanczos-3, bicubic or bilinear, computed in JavaScript in your browser. There is no neural network and no external service, which means nothing is uploaded, it works offline once loaded, and the same input always produces exactly the same output rather than a generated guess.
Lanczos-3, which is the default. It uses a windowed sinc function over a six-by-six neighbourhood and preserves edges and fine detail better than the alternatives. Bicubic is a close second at lower cost, and bilinear is the fastest but visibly softest. The trade-off is processing time, which rises with filter width.
No. Interpolation estimates new pixels from existing ones and cannot recover detail the camera never captured. A blurry source becomes a larger blurry image with smoother edges. It also magnifies compression artefacts and noise, so always start from the cleanest, least compressed original you have available.
Because output area grows with the square of the scale factor. A 2x upscale produces four times as many pixels as the source, while 4x produces sixteen times as many, and every pixel requires a weighted sum over its neighbourhood. Under Lanczos-3 that is around 36 source samples per output pixel per channel.
PNG for screenshots, logos, line art and anything needing transparency, because it is lossless and will not add compression damage. For photographs, JPEG or WebP at 85 to 95 percent quality keeps the file manageable. WebP is generally smaller than JPEG at equivalent quality and is supported by all current browsers.
Lanczos, bicubic and bilinear resampling compared
Upscaling means inventing pixels that were never captured, and the algorithm decides how those pixels are estimated from their neighbours. All three methods here are interpolation filters that compute each output pixel as a weighted average of nearby source pixels, differing in how many neighbours they consider and how the weights are shaped.
Bilinear takes the four nearest source pixels and blends them linearly by distance. It is the cheapest to compute and produces smooth but noticeably soft results, because the linear weighting blurs edges.
Bicubic uses a four-by-four neighbourhood, sixteen source pixels, with the Keys cubic weighting function. The cubic curve holds edges better than a straight line, giving a sharper result than bilinear at moderately higher cost.
Lanczos-3 uses a windowed sinc function with a radius of three, drawing on a six-by-six neighbourhood of thirty-six source pixels. Sinc is the theoretically ideal reconstruction filter, and windowing makes it practical. It preserves fine detail and edge definition best, at the highest cost, which is why it is the default and also the slowest.
Resampling is applied separately in the horizontal and vertical directions, and the result is drawn to a canvas and encoded in your chosen format.
Bilinear takes the four nearest source pixels and blends them linearly by distance. It is the cheapest to compute and produces smooth but noticeably soft results, because the linear weighting blurs edges.
Bicubic uses a four-by-four neighbourhood, sixteen source pixels, with the Keys cubic weighting function. The cubic curve holds edges better than a straight line, giving a sharper result than bilinear at moderately higher cost.
Lanczos-3 uses a windowed sinc function with a radius of three, drawing on a six-by-six neighbourhood of thirty-six source pixels. Sinc is the theoretically ideal reconstruction filter, and windowing makes it practical. It preserves fine detail and edge definition best, at the highest cost, which is why it is the default and also the slowest.
Resampling is applied separately in the horizontal and vertical directions, and the result is drawn to a canvas and encoded in your chosen format.
Worked example: a 640 by 480 image at 3x
Take a 640 by 480 screenshot and select the 3x scale factor. The output dimensions preview shows 1920 by 1440, since each dimension is multiplied by three.
The pixel count is where the cost becomes clear. The source holds 307,200 pixels. The output holds 2,764,800, which is nine times as many, because area scales with the square of the linear factor. Every one of those 2.76 million output pixels must be computed, and under Lanczos-3 each one is a weighted sum over a six-by-six neighbourhood, so roughly 36 multiply-and-add operations per channel per pixel. That is why the reported processing time rises steeply with the scale factor, and why 4x on a large photograph takes appreciably longer than 2x on a small one.
Only about 11 percent of the output pixels correspond to an original sample; the rest are interpolated estimates. This is the fundamental limit: the result holds no more real detail than the 640 by 480 source did. A good filter buys a cleaner, less blocky presentation of existing detail, not new information.
Saved as PNG at 1920 by 1440, the file is several times larger than the original despite carrying no extra detail.
The pixel count is where the cost becomes clear. The source holds 307,200 pixels. The output holds 2,764,800, which is nine times as many, because area scales with the square of the linear factor. Every one of those 2.76 million output pixels must be computed, and under Lanczos-3 each one is a weighted sum over a six-by-six neighbourhood, so roughly 36 multiply-and-add operations per channel per pixel. That is why the reported processing time rises steeply with the scale factor, and why 4x on a large photograph takes appreciably longer than 2x on a small one.
Only about 11 percent of the output pixels correspond to an original sample; the rest are interpolated estimates. This is the fundamental limit: the result holds no more real detail than the 640 by 480 source did. A good filter buys a cleaner, less blocky presentation of existing detail, not new information.
Saved as PNG at 1920 by 1440, the file is several times larger than the original despite carrying no extra detail.
What upscaling can and cannot recover
Set expectations correctly. Interpolation cannot reconstruct detail that was never recorded. If a face is eight pixels wide in the source, no resampling filter will produce recognisable features at 4x. Text that is illegible in the original stays illegible, just larger and smoother. This tool is not an AI upscaler and makes no attempt to hallucinate plausible detail; it applies a deterministic mathematical filter, which is why it gives the same result every time.
Where it works well is enlarging already-clean sources: line art, logos, diagrams, screenshots at moderate factors, and reasonably sharp photographs that simply need to fill a larger frame. Use the compare view at full size rather than judging from a thumbnail, since interpolation differences are invisible when the image is scaled down for display.
Existing flaws are magnified too. JPEG artefacts, sensor noise and moiré all become larger and more obvious, so start from the cleanest source available and prefer an untouched original.
On output format, choose PNG for graphics, screenshots and anything with flat colour or transparency, since it is lossless. For photographs, JPEG or WebP at 85 to 95 percent avoids adding fresh compression damage on top of the enlargement.
Where it works well is enlarging already-clean sources: line art, logos, diagrams, screenshots at moderate factors, and reasonably sharp photographs that simply need to fill a larger frame. Use the compare view at full size rather than judging from a thumbnail, since interpolation differences are invisible when the image is scaled down for display.
Existing flaws are magnified too. JPEG artefacts, sensor noise and moiré all become larger and more obvious, so start from the cleanest source available and prefer an untouched original.
On output format, choose PNG for graphics, screenshots and anything with flat colour or transparency, since it is lossless. For photographs, JPEG or WebP at 85 to 95 percent avoids adding fresh compression damage on top of the enlargement.