JPG Compressor — Compress JPG Online (Free)

Why compress JPG?

Image weight is the biggest single contributor to slow page loads. A modern phone shoots 4–8 MB photos out of the box, and even a single such image embedded in a web page delays Largest Contentful Paint significantly. Google's Core Web Vitals — the speed signals that affect search ranking — punish heavy images directly. Compressing JPGs to 100–400 KB instead of multi-MB pays off twice: pages load faster for users, and search engines reward you for it.

Even outside the web, smaller JPGs matter. Email attachments under 25 MB total are a soft rule for most providers. Disk space on phones, on iCloud, on Google Drive — every byte counts. A single compress-and-replace pass typically cuts a photo library's footprint by 70–85% with no visible loss.

How JPG compression works

JPG is a "lossy" format — it intentionally discards data the human eye is least likely to notice in exchange for dramatic file-size reductions. The encoder splits the image into 8×8 pixel blocks, transforms each block into frequency components using a Discrete Cosine Transform, then aggressively quantises the high-frequency components (the fine details we don't notice much) while preserving the low-frequency ones (the broad shapes and tones we do notice).

The "quality" setting controls how aggressively the high-frequency components are discarded. Lower quality = more discarded = smaller file = more visible artefacts at extreme settings. Higher quality = less discarded = larger file = more faithful to the original. There's a sweet spot for every image, typically between 70–85, where the file is dramatically smaller and you'd struggle to tell the result apart from the original.

Quality settings — what each range gives you

• 95–100 — near-lossless. File 20–40% smaller than the source. Use for archival masters.
• 85–95 — visually identical for most photos. File 40–60% smaller. Use for hero images and high-quality web display.
• 75–85 — the sweet spot. File 60–80% smaller. Imperceptible loss on photos. Default for general web use.
• 60–75 — visible only on close inspection. File 70–85% smaller. Great for thumbnails and below-the-fold images.
• 45–60 — obvious compression. Use only when size matters more than quality (placeholders, very small thumbnails).
• Below 45 — heavily compressed, blocky. Use sparingly.

Bonus: resize + compress for maximum savings

If your source JPG is much larger than its display size, the single biggest size win is resizing — not compressing. A 4000-pixel-wide phone photo displayed at 800 px wide is carrying ~25× more data than necessary. Untick "Keep original dimensions" and pull the max-width slider to your display size; the compressed output will be a fraction of the source's weight without any visible quality loss at the intended size.

A common recipe for hero web images: max width 1920, quality 80. Result is usually 150–300 KB from a 4 MB source.

How this compressor works

Your JPG is decoded into a hidden canvas, optionally resized, then re-encoded as JPG at your chosen quality. The whole pipeline runs in your browser using canvas.toBlob('image/jpeg', quality). Nothing is uploaded. The encoder is the same one your browser uses everywhere else; output quality matches what you'd see in Squoosh or any other browser-based tool.

Tips and best practice

• Don't re-compress an already-compressed JPG. Each save adds new artefacts on top of the old ones. Start from the highest-quality version you have.
• For photographs, quality 80 is almost always the right answer. If you can see a difference, try 85. If 80 looks too soft, your source was probably already heavily compressed.
• Resize first, compress second. The savings compound.
• Use WebP for the web instead. The PNG-to-WebP tool produces files 25–35% smaller than JPG at the same visual quality.
• Strip EXIF. Re-encoding through canvas automatically removes location, camera and personal metadata.

JPG compression FAQs

Related JPG tools