Free JPG to PNG Converter

What "JPG to PNG" actually does (and doesn't do)

JPEG and PNG are mirror opposites on the lossless-versus-lossy axis. JPEG (ISO/IEC 10918-1, 1992) uses DCT-based lossy compression: the encoder discards image data the human eye is unlikely to notice in exchange for much smaller files. PNG (IETF RFC 2083, 1996; W3C 3rd Edition, 2023) uses DEFLATE lossless compression: every pixel is preserved bit-exact. Converting JPEG to PNG goes against the file-size grain (PNG output is typically 3-5x larger than the source JPEG for photographic content) and crucially does not recover any of the detail the original JPEG threw away. The PNG faithfully preserves the JPEG's already-lossy pixels: every blocking artifact, chroma bleed and edge halo is baked into the output exactly as it appeared in the input.

So why convert at all? Because the conversion freezes the lossy degradation at its current level and gives you a format that preserves any further work losslessly. If you plan to edit a JPEG photo (crop, retouch, colour-correct) and save several times, each JPEG save introduces fresh DCT quantization error and the image degrades visibly after three or four passes. Convert once to PNG and the subsequent edits round-trip without compounding loss. Other valid reasons: a destination platform requires PNG specifically, you want to add transparency later (PNG supports alpha; the conversion output is fully opaque but the format will accept future alpha edits), or you want a more conservative long-term archival format.

What the conversion does not do: it does not "increase quality," "remove JPEG artifacts," or "restore lost detail." Information theory is unforgiving here. The bits the JPEG encoder discarded are gone. The PNG cannot synthesise them back. If your JPEG has visible blocking or ringing, the PNG will have the same blocking and ringing, just in a larger file. The only way to truly recover detail from a low-quality JPEG is AI super-resolution (Real-ESRGAN, Topaz Gigapixel, Adobe Super Resolution), which hallucinates plausible detail using neural networks; that's a different operation from format conversion.

How this tool works under the hood

The conversion uses the HTML5 Canvas 2D API and JSZip (Stuart Knightley, MIT/GPL dual licence) for multi-file packaging. No external decoder is needed because every browser decodes JPEG natively. When you drop a JPEG, the File API hands the bytes to a new HTMLImageElement; the browser's built-in JPEG decoder performs the standard inverse-DCT, inverse-quantization, YCbCr-to-RGB colour conversion, and produces a raw RGBA pixel buffer with alpha set to fully opaque (JPEG has no alpha to populate).

A fresh in-memory <canvas> is sized to the image and the decoded pixels are painted on with ctx.drawImage(). Then canvas.toBlob('image/png') invokes the browser's built-in PNG encoder. The encoder runs DEFLATE (LZ77 + Huffman) over the RGBA byte stream and writes a standards-compliant PNG file. There is no quality argument: PNG is always lossless, so the encoder's only knob is the DEFLATE effort level, which browsers fix at a sensible default.

For single files the output Blob becomes a downloadable object URL. For multi-file batches, JSZip packages every PNG output into a single ZIP archive that the browser serves as one download. Nothing leaves the tab. The only network traffic is the initial page load and the JSZip library (~95 KB on first visit, cached afterwards). Open DevTools' Network tab while converting: no requests carry image data. Switch the browser to airplane mode after the page loads and the converter keeps working on local JPEGs.

A brief history of JPEG and PNG

• JPEG, 1992. ITU-T T.81 / ISO/IEC 10918-1, standardised by the Joint Photographic Experts Group (formed 1986). 8x8 DCT blocks, YCbCr colour with optional chroma subsampling, quantization tables tuned for human vision. Lossy, no alpha channel. The format that made the photo-rich web possible.
• The LZW patent crisis, 1994-1995. Unisys began asserting patent claims over LZW compression used by GIF, demanding licence fees from any software that creates GIF files. The web community organised around a patent-free replacement for the lossless-with-alpha use case GIF had served.
• PNG 1.0, January 1996. Thomas Boutell, Mark Adler and a CompuServe-led working group published PNG at the IETF as RFC 2083. DEFLATE compression (the same LZ77+Huffman algorithm as zlib and ZIP, also patent-free) replaced LZW. Full 8-bit alpha channel, always lossless, royalty-free. W3C Recommendation followed in 1997.
• Two formats, two jobs, late 1990s onward. JPEG became the de facto format for photographs; PNG became the standard for graphics, logos, screenshots, anything requiring transparency or lossless rendering. The cross-conversion has been a daily workflow ever since: from JPEG to PNG when editing or platform requirements demand it, from PNG to JPEG to shrink photographic content.
• ISO standardisation and APNG, 2003-2010. PNG became ISO/IEC 15948 in 2003. Mozilla developed APNG (animated PNG) in 2004 as a Firefox extension; it gained Safari support in 2014 and Chrome in 2017, but the W3C PNG group only formally adopted it in the 3rd Edition in 2023.
• PNG 3rd Edition, 2023. W3C published the 3rd Edition of PNG with formal HDR support (cICP chunk), APNG animation, and standardised EXIF metadata blocks (eXIf chunk). Over thirty years after JPEG, both formats remain the dominant image formats on the web, with WebP and AVIF gaining ground but not displacing them.

How It Works

1. Upload JPGs: Drop or select one or multiple JPG/JPEG files to convert.
2. Convert: Click "Convert to PNG" to start the conversion process. Processing happens instantly in your browser.
3. Download: Single images download directly; multiple files are packaged into a ZIP archive for easy batch downloading.

Why Convert JPG to PNG?

PNG is a lossless format that preserves all image data without quality loss. Unlike JPG, PNG supports transparency and produces crisp, detailed images ideal for graphics, logos, and web design. Convert to PNG when you need transparency support, lossless quality for editing, or archival storage. Note: PNG files are typically larger than JPG, but offer better quality and flexibility.

Features

• Batch conversion: Convert multiple JPG files at once. Single images download directly, multiple images are zipped.
• Lossless quality: PNG preserves all image data without compression artifacts or quality loss.
• Transparency support: PNG supports transparent backgrounds, perfect for graphics and design work.
• Privacy: All conversion happens locally in your browser. No uploads to any server.
• Fast processing: Instant conversion with no waiting or queues.

Real-world JPG-to-PNG workflows

• Preparing a JPEG for further editing. A photo arrives as JPEG. Before opening it in Photoshop, Affinity Photo or GIMP for retouching, crop or colour-correct work, convert to PNG. Each subsequent save during the edit will be lossless. If you stayed in JPEG and saved several times during the workflow, the image would visibly degrade after three or four passes (compounding DCT quantization error). The PNG conversion freezes the existing lossy state and prevents further degradation.
• Submission to a service that requires PNG. Some photo-print services, government portals, academic journal submission systems, design contest portals, and legacy CMS installations accept PNG and reject JPEG outright. Convert before upload to avoid the round-trip when the submission is rejected with a cryptic "unsupported format" error.
• Adding text or sharp graphics over a photo. A magazine cover, an infographic, a meme overlay. JPEG's DCT compression produces halos around sharp colour transitions: text laid over a photo and saved as JPEG looks visibly degraded along every letter edge. PNG handles the mixed photo-plus-graphics content cleanly. Convert the photo to PNG first, do the overlay work, and save the combined image as PNG.
• Lossless archival of edited photos. Once you've finished colour-correcting or retouching a JPEG source, saving the result as PNG means future re-saves won't introduce additional quantization error. If you'll likely reopen and modify the image later (slightly different crop, additional adjustments), the PNG archive avoids generation loss.
• Preparing for transparency edits. The output PNG is fully opaque (the JPEG source had no alpha to populate). But PNG supports an alpha channel, so subsequent edits can cut out a background, add semi-transparent halos, or composite the image over other graphics. Doing the same workflow in JPEG would require a final PNG conversion at the end anyway, so converting upfront keeps every step lossless.
• Web and UI design where pixel-perfect rendering matters. Design portfolios, UI mockups, screenshots that mix photographic and typographic content, app icons derived from photos. PNG is the standard format for these because JPEG halos at sharp colour edges are jarring in design contexts. Converting JPEG source assets to PNG once at the start of a design project keeps the rest of the workflow consistent.

Common pitfalls and what they mean

• The PNG does not recover lost JPEG detail. This is the most common misconception. The JPEG encoder discarded image data; information theory says you cannot reconstruct what was discarded. The PNG preserves the JPEG's already-lossy pixels exactly, including every blocking artifact and edge halo. If you need to truly recover detail, that requires AI super-resolution (Real-ESRGAN, Topaz Gigapixel, Adobe Super Resolution), which hallucinates plausible detail using a neural network. Format conversion alone cannot do this.
• The file size always grows. For typical photographic content the PNG is 3-5x larger than the source JPEG. Sometimes 10x for very compressed JPEGs. PNG's DEFLATE compression is no match for JPEG's DCT for continuous-tone images; PNG cannot drop any of the JPEG noise patterns to save space because it has to be lossless. If file size matters more than lossless quality, do not convert to PNG.
• JPEG artifacts are preserved exactly. DCT blocking in flat areas, chroma bleed at colour boundaries, ringing halos near hard edges, mosquito noise around text. All of these are now in the PNG at full fidelity. The conversion does not soften, deblock, or denoise them. They will be visible in print and in any pixel-peeping comparison. The PNG just freezes them in place.
• Avoid JPEG → PNG → JPEG loops. A common mistake: convert JPEG to PNG for editing, edit, then save the result as JPEG and repeat tomorrow. The intermediate PNG doesn't help: each round-trip to JPEG re-runs the lossy DCT quantization. To benefit from PNG's lossless saves, do all the editing in PNG and only re-encode to JPEG at the very end if needed.
• EXIF and ICC metadata may be stripped. Canvas-based re-encoding typically discards EXIF, IPTC, XMP metadata blocks (camera, exposure, GPS, copyright) and the embedded ICC colour profile. For ordinary screen use this is invisible; for archival workflows where the metadata matters, use a metadata-aware desktop tool like ImageMagick with explicit options or jpegtran-equivalent metadata-preserving paths.
• Very large JPEGs need significant memory. Decoding a JPEG to RGBA pixels needs RAM proportional to its dimensions: a 24-megapixel photo (6000x4000) needs about 96 MB for the source pixel buffer, plus a separate buffer for the canvas, plus the PNG encoder workspace. Mobile devices with 1 to 2 GB of RAM available to the browser may have the tab terminated by the OS before the encode completes.

Privacy: images never leave your device

Every cloud-based JPG-to-PNG converter (CloudConvert, Convertio, iLoveIMG, FreeConvert, Aspose, and the dozens of "jpg to png online" services) uploads your file to the operator's server, runs the conversion, and returns the PNG as a download. JPEGs from phones and cameras routinely embed EXIF metadata: GPS coordinates of where the photo was taken (often precise to a few metres), camera and lens model, capture date and time, even small thumbnail previews. All of that goes to the operator. Most operators publish privacy policies committing to delete uploads within an hour or two and to encrypt in transit, and the larger ones hold ISO/IEC 27001 certification. They have strong commercial reasons to honour those policies. But "deleted within an hour" is not "never seen." During that hour the photo content sits in operator infrastructure, accessible to any process or person with appropriate permissions, and visible in logs and backups according to whatever retention policy applies.

This converter never uploads anything. The entire pipeline (file pick, JPEG decode via the browser's built-in decoder, Canvas paint, PNG encode, optional ZIP packaging, download) runs inside your browser tab using JavaScript and the HTML5 Canvas API. No upload, no network request carrying image data, no log entry. You can verify by opening the browser developer tools to the Network tab before converting: no request fires with image content. The only network traffic is the initial page load and the small JSZip library (~95 KB), fetched once from the CDN on first visit and cached. Switch the browser to airplane mode after the page loads and the converter keeps working on local JPEGs.

When another tool is the right choice

• You want to recover lost detail or remove JPEG artifacts. Format conversion cannot do this. Use AI super-resolution: Real-ESRGAN (open source, runs locally via ChaiNNer or the command line), waifu2x (open source, free web demos), Topaz Gigapixel AI (commercial desktop), or Adobe Super Resolution inside Camera Raw or Lightroom. These models hallucinate plausible detail using neural networks trained on millions of similar images.
• File-size minimisation matters more than lossless quality. Don't convert to PNG; PNG will be 3-5x larger. If the JPEG is acceptable quality, stay in JPEG. If you need a smaller file, run the JPEG through the Image Compressor at a lower quality setting, or convert to WebP via the Image Converter (WebP-lossy is typically 25-35% smaller than JPEG at the same quality).
• Batch automation across hundreds of files. Use Sharp in Node.js (the canonical server-side image library built on libvips: sharp(buf).png().toBuffer()), ImageMagick on any shell (magick input.jpg output.png), or Pillow in Python (Image.open(p).save(out, 'PNG')). CLI tools handle thousands of files without browser memory limits and run from CI jobs, deploy hooks or cron tasks.
• Print-grade workflows with metadata and colour-profile preservation. Use Photoshop's Save As PNG (preserves ICC profile and metadata), Affinity Photo, or RawTherapee. Canvas-based browser conversion may strip the embedded ICC profile and EXIF blocks, which is fine for screen use but not for print prep or workflows that depend on capture metadata.

Frequently Asked Questions

Will my PNG be larger than the original JPG?

Likely yes. PNG uses lossless compression while JPG uses lossy compression, which typically results in smaller file sizes. However, PNG provides better quality and supports transparency. Choose PNG when quality and flexibility are more important than file size.

Can I convert multiple files at once?

Yes. Upload multiple JPG files and they will all be converted with the same settings. Single files download directly, multiple files are automatically packaged into a ZIP archive.

What's the file size limit?

Each JPG can be up to 50 MB. Total conversion size depends on your browser's available memory, but typically you can convert files totaling several hundred MB.

Are my images uploaded to a server?

No. All conversion happens locally in your browser using the Canvas API. Your images never leave your device, ensuring complete privacy and security.

Can I convert JPG to PNG on mobile?

Yes. This tool works on desktop, tablet, and mobile browsers. Just tap to select files and convert.

What about JPG vs PNG quality?

PNG conversion preserves the image quality of the original JPG. Since JPG is already lossy, converting to PNG won't restore lost detail, but it prevents further quality loss if you need to edit the image.

More frequently asked questions

Will converting to PNG remove the JPEG artifacts I can see?

No. JPEG artifacts (DCT blocking in flat regions, halos and ringing at sharp edges, chroma bleed at colour transitions) are part of the pixel data after the JPEG encoder has done its work. PNG is a lossless format: it preserves the pixel data exactly as it is, artifacts and all. The output PNG will look identical to the source JPEG at every zoom level. If you need to soften or remove visible JPEG artifacts, the only options are AI denoising/super-resolution tools or careful manual retouching; format conversion cannot do this.

Why is the PNG so much bigger?

Because PNG's DEFLATE compression is much less effective than JPEG's DCT for photographic content. JPEG can compress smooth gradients and continuous-tone regions extremely efficiently by discarding high-frequency information the eye won't notice. PNG must preserve every pixel exactly, so it has to encode the full noise pattern of the photograph, which DEFLATE can do only mildly. For typical photos, PNG is 3-5x larger than the equivalent JPEG at quality 85; sometimes 10x for very compressed JPEGs. This is expected behaviour, not a bug in the tool.

Can the resulting PNG have transparency?

Not automatically, because the source JPEG has no alpha channel to populate. The output PNG is fully opaque. But the PNG format does support an 8-bit alpha channel, so you can open the PNG in a later editing step (Photoshop, GIMP, Photopea, Pixelmator) and add transparency: cut out a background, add a semi-transparent halo, composite over other graphics. The conversion gives you a format that accepts those future edits without another format change.

Does this tool work offline?

Yes. The JPEG decoder and PNG encoder are both built into every browser; no external library is downloaded for them. The only library this tool loads is JSZip (~95 KB) for packaging multi-file batches into a ZIP archive, fetched once from the CDN on first load and cached. Subsequent visits work entirely offline. You can verify by enabling airplane mode after opening the page once and converting a local JPG file.

Will my EXIF metadata (camera, GPS, date) survive?

Usually no. Canvas-based re-encoding strips EXIF, IPTC and XMP metadata blocks (camera model, exposure settings, GPS coordinates, copyright tags) along with the embedded ICC colour profile. The output PNG carries pixel data only. For ordinary screen sharing this is usually a privacy win (GPS coordinates and device serial numbers won't leak). For archival workflows where you need to preserve the capture metadata, use a metadata-aware desktop tool: ImageMagick with the right options, ExifTool, or Photoshop's Save As, all of which can preserve EXIF data across the conversion.

Is there a desktop or command-line equivalent?

Several. For batch automation, sharp in Node.js is the standard server-side library: sharp(buf).png().toBuffer(). ImageMagick on any shell: magick input.jpg output.png. Python's Pillow: Image.open(p).save(out, 'PNG'). For one-off interactive work like this tool, Squoosh (Google Chrome Labs, entirely client-side) is a closer alternative and supports more output formats including AVIF. Photoshop, Affinity Photo and Preview on macOS handle the desktop GUI case and preserve metadata more reliably than browser-based conversion.

Related Tools