Převodník obrázků

A free online image converter that runs in your browser

This image converter changes your pictures between JPG, PNG and WebP — and reads HEIC, GIF, BMP and AVIF where your browser supports them — without ever uploading a file. Every image is decoded and re-encoded locally on a <canvas>, so there are no upload limits, no sign-up, and no waiting in a free-tier queue. The tool works in Chrome, Firefox, Edge and Safari with no install, no extension and no account.

Whether you need an image converter to JPG for sharing a screenshot, an image converter to PNG to keep transparency, or to switch either to WebP for a faster website, you can do it in a couple of clicks. The batch mode handles a folder of photos as easily as a single file, and a before-and-after readout shows exactly how much each file shrank.

Convert to JPG, PNG or WebP — one image or a whole batch

Drop in a single picture or a stack of them. Set the output format, quality and options once, then click Convert all and download each file individually or all of them together as a single ZIP. Common conversions like PNG to JPG, JPG to PNG, WebP to PNG, PNG to WebP and JPG to WebP all work the same way. The per-file card shows original size, converted size and the percentage reduction so you know the conversion was worth it.

Because the work happens on your device, there is no per-file size cap beyond your computer's memory. A folder of 50 photos converts as fast as your CPU allows — no server queue, no daily conversion limit, and no paid plan required to unlock batch mode.

Black-and-white, resize and a quality slider built in

This is more than a format switch. Flip the black and white toggle to desaturate to true grayscale during conversion — useful for print layouts, vintage edits or simply shrinking a photo by removing color data. Use resize to scale down by percentage or cap the longest side to a pixel width — helpful for trimming a 6000-pixel phone photo down to 1280 pixels for a web page. For lossy formats, the quality slider lets you trade a little sharpness for a much smaller file; quality 80–90 is the sweet spot for most photos.

Converting a transparent image to JPG? Pick a background fill color so the transparent areas blend into whatever surface the image will sit on, instead of turning black or white unexpectedly. Strip EXIF metadata to remove embedded camera model, GPS coordinates and shooting settings before sharing — especially useful for personal photos or screenshots that capture sensitive screen content.

JPG vs PNG vs WebP — which should you choose?

Use JPG for photographs and anywhere file size matters most; it's universally supported but has no transparency channel. Use PNG for logos, icons, screenshots and anything with sharp edges or transparency, since it's lossless and every pixel you put in is exactly what you get back. Use WebP when you want the best of both: it handles photos and transparent graphics and is typically 25–35% smaller than JPG at the same visual quality — which is why Google's PageSpeed Insights and Core Web Vitals tools recommend it for modern sites.

AVIF goes further on compression but encoding support is less consistent across tools and platforms. If you're sending to email or printing, JPG remains the safe, universally compatible choice. For social platforms that re-compress uploads anyway (Instagram, Twitter/X), WebP or a high-quality JPG gives the best starting point.

How this compares to CloudConvert, Convertio and Squoosh

Server-based tools are the main alternative — and they come with real tradeoffs. Convertio.co limits free users to files under 100 MB and 10 conversions per day; going over requires a paid plan starting at $9.99/month. CloudConvert.com bills by the conversion minute and gives you 25 free minutes per day before charging — one heavy batch can use that up. Both send every file to their servers, adding latency and a privacy concern for personal or sensitive images.

Squoosh (Google's browser-based compressor) is free and private but handles only one image at a time with no ZIP download or batch. This converter batches unlimited files, keeps everything on your device, and adds resize, grayscale and background-fill controls that Squoosh doesn't offer. No account required, no watermark on any tier, and no limit on how many times you use it in a day.

Private by design — nothing leaves your device

Privacy is the clearest difference between browser-based and server-based converters. When you upload to Convertio or CloudConvert, your photos — and any embedded location or camera data — travel to a third-party server and are processed there. Here, every image is decoded and re-encoded inside the browser tab using the native <canvas> API. Nothing is transmitted, no copy is stored, and no network request is made for the conversion itself.

That makes this tool safe for personal photos, medical images, ID documents, unreleased product designs and confidential screenshots. Once the page has loaded, it works offline too — no connection required to convert. Bookmark this image converter for any time you need to change a format, shrink a file or strip it to black and white without sending anything off your device.

A brief history of image formats: from GIF to AVIF

GIF (Graphics Interchange Format) arrived in 1987 from CompuServe and established early web graphics — but its 256-color palette limit made it unsuitable for photographs. The underlying LZW compression algorithm was patented by Unisys, and those patents held back open-source tooling until they finally expired in 2003. JPEG changed everything: standardized in 1992 by the ITU-T and ISO, it introduced lossy DCT compression specifically designed for photographs, and it remained the dominant web image format for nearly three decades because it delivered excellent quality at small file sizes for natural imagery.

PNG (Portable Network Graphics) launched in 1995 specifically as a patent-free alternative to GIF. It supports true color (24-bit RGB) and a full alpha transparency channel, and uses lossless DEFLATE compression — making it the right format for logos, icons and screenshots rather than photographs. WebP, released by Google in 2010, was engineered to replace both JPEG and PNG simultaneously: its lossy mode is 25–34% smaller than JPEG at equivalent quality, and its lossless mode is approximately 26% smaller than PNG, while also supporting transparency. AVIF (AV1 Image File Format), finalized in 2019 and built on the same codec as the AV1 video standard, pushes compression further still — achieving roughly 50% size reduction compared to JPEG — and gained broad browser support across Chrome, Firefox and Safari by 2023–2024. JPEG XL, standardized in 2021–2022, offers a unique capability: lossless recompression of existing JPEG files with no quality loss, along with very high compression efficiency for new images, but browser adoption has been limited — Chrome removed its trial support in 2022, citing insufficient developer interest.

Understanding this lineage matters when choosing a format. GIF survives only for simple animations; JPEG remains the universal fallback for photographs on platforms that reject WebP; PNG holds the line for graphics with transparency; WebP is the pragmatic modern default; and AVIF is the high-efficiency choice for teams that control their serving stack. Each format reflects the technical constraints and priorities of its era.

Color depth: what 8-bit, 24-bit and 32-bit actually mean

Color depth describes how many distinct colors a single pixel can represent. At 1-bit, each pixel is either black or white — no greys — which is the correct mode for QR codes, barcodes and scanned text documents because it produces the smallest possible file with sharp, machine-readable edges. 8-bit indexed color maps each pixel to one of 256 entries in a color palette; this is how GIF works, and it is also supported by palette-based PNG. For simple graphics with few colors — flat-color icons, pixel art, simple charts — 8-bit indexed color can be smaller than full-color formats, but it falls apart on photographs where smooth gradients require far more than 256 colors. 8-bit greyscale encodes 256 shades of grey and is the correct mode for black-and-white photographs and medical imaging where no color information is needed.

The web standard for color photographs is 24-bit RGB: 8 bits per channel across three channels (red, green, blue) gives 16.7 million possible colors. This is the native mode of JPEG and is what your browser assumes when it renders a standard image. Adding an alpha transparency channel produces 32-bit RGBA — the 8-bit alpha encodes 256 levels of opacity from fully transparent to fully opaque, which is how PNG preserves soft shadows, anti-aliased edges and layered graphics. Professional photography workflows often work at 16 bits per channel (48-bit color), which RAW files from cameras typically capture at 12–14 bits per channel. The extra bit depth provides far more editing headroom before clipping — you can recover detail from shadows or highlights without introducing banding or posterization in ways that 8-bit editing cannot match.

Modern mobile and display technology is pushing into HDR color depths. AVIF and HEIC both support 10-bit per channel encoding and wide color gamuts such as Display P3, which Apple uses across iPhone cameras and Retina displays. A 10-bit P3 image contains color information that a standard 8-bit sRGB JPEG simply cannot store. When this converter processes images via the browser Canvas API, it operates in the sRGB color space at 8 bits per channel — which means HDR or wide-gamut source images are tone-mapped and gamut-clipped to sRGB in the process. For workflows where HDR accuracy matters, dedicated software such as Lightroom or darktable is the appropriate tool.

Color profiles and the sRGB web requirement

ICC color profiles (International Color Consortium) are metadata blocks embedded in image files that tell color-managed software exactly how to interpret the numerical RGB values in the file. The three profiles you encounter most often are sRGB (the standard for consumer displays and the web), Adobe RGB (a wider gamut used in professional print photography) and ProPhoto RGB (an extremely wide gamut used by some RAW processing pipelines). The same numerical value — say, R=200, G=50, B=50 — looks like a different shade of red depending on which profile is declared.

The critical web publishing rule is: always convert to sRGB before uploading images to a website. The sRGB color space covers approximately 35% of the colors visible to the human eye, which is the range a typical consumer monitor can display. Adobe RGB covers roughly 50% of the visible gamut, and it is used by professional photographers shooting in JPEG+Adobe RGB or RAW because the wider gamut gives more room to adjust saturation and hue in post-processing before the image is flattened to sRGB for delivery. The problem is that browsers default to assuming sRGB. An Adobe RGB JPEG displayed without color management — which is the behavior of browsers that ignore the embedded profile — renders every color as if it were sRGB, making the image appear visibly washed out and desaturated. Converting to sRGB before publishing eliminates this mismatch and ensures consistent appearance across all browsers and devices.

A related concern is EXIF metadata: the block of data JPEG stores alongside pixel values that records camera make and model, lens, shutter speed, aperture, ISO, the date the photo was taken, and — if geotagging was enabled — precise GPS coordinates. Some image converters and compression tools silently strip EXIF, which can destroy embedded copyright and creator information (stored in the EXIF IPTC fields) or remove location data you intended to keep. Others faithfully preserve it, which can inadvertently expose GPS coordinates in images you share publicly. This converter strips EXIF during Canvas re-encoding by default — the browser Canvas API does not propagate EXIF from the source — which is safe for web publishing but means you should keep your originals if the metadata has archival value.

How the Canvas API performs client-side image conversion

The browser's HTML5 Canvas API is the engine behind every client-side image converter, including this one. The conversion pipeline is straightforward: an <img> element (or a Blob URL created from the file) is drawn onto an off-screen <canvas> using drawImage(), and then the canvas is serialized back into a new image file using either canvas.toDataURL('image/webp', 0.85) for an inline Base64 string or the more memory-efficient canvas.toBlob(callback, 'image/jpeg', 0.90) which produces a raw Blob the browser can hand directly to a download link. The quality parameter in both methods maps to the lossy compression level — 0.0 is maximum compression, 1.0 is minimum compression — and it only applies to JPEG and lossy WebP; PNG output is always lossless regardless of the value passed.

The MIME types a browser supports for output encoding vary by engine. Every modern browser encodes image/jpeg and image/png. image/webp output is supported in Chromium-based browsers and Firefox 96+, and in Safari 16+. image/avif encoding via canvas.toBlob() is not yet universally available — Chromium supports it, but Safari's implementation is incomplete and Firefox support was added only in recent versions. For reliable AVIF encoding at scale, server-side pipelines using libavif or ffmpeg with the libaom encoder are more dependable than relying on the Canvas API. An alternative browser pipeline for high-throughput work is fetch() + createImageBitmap(), which decodes the image into a GPU-accelerated bitmap without ever creating a DOM <img> element, reducing memory pressure when processing many large files in sequence.

There are three inherent limitations of the Canvas conversion approach that every browser-based tool shares. First, the Canvas operates in the sRGB color space at 8 bits per channel, so any wide-gamut or HDR source data is silently clipped and tone-mapped during drawImage(). Second, all EXIF and ICC metadata is stripped — the canvas has no mechanism to propagate it to the output blob, so the converted file always emerges without camera settings, GPS data or an embedded color profile (browsers implicitly write sRGB). Third, animated images require frame-by-frame processing: a GIF or animated WebP has no native animation support via Canvas, so converting to an animated output format requires decoding each frame individually, drawing it, capturing the blob, then reassembling the sequence — a significantly more involved process than a single-frame still conversion.