Supported formats & honest limitations

Works with: JPG/JPEG, PNG, and WebP input, up to 25 MB per file. Output can be JPG, PNG, or WebP, and dimensions are always preserved — compression never stretches or crops your image.

Lossy vs lossless: JPG and WebP compression is lossy — it permanently discards some image detail to save space. At moderate quality (75–85%) the difference is usually invisible in photos, but fine text, logos, and sharp edges can develop fuzziness (“artifacts”). PNG output is lossless, so it keeps every pixel — but re-saving a PNG rarely makes it much smaller, and converting a photo to PNG usually makes the file bigger. We show you the true result either way, including when a file grows.

Browser limits: because everything runs on your device, very large photos on older phones can be slow, and animated GIFs are not supported (a GIF would lose its animation). For a file that must hit an exact byte limit for a government or exam form, use the target-size option and always keep your original as a backup.

A practical example

A phone photo saved as a 4.2 MB JPG typically compresses to roughly 600 KB–1 MB at 80% quality as JPG, and often 20–40% smaller again as WebP — with no visible change at normal viewing sizes. Exact results depend on the image: detailed scenes (grass, crowds, confetti) compress less than smooth ones (skies, studio backdrops). That is why this tool reports your actual numbers instead of promising a percentage.

Supported formats & honest limitations

Works with: JPG/JPEG, PNG, and WebP input up to 25 MB; output as JPG, PNG, or WebP. Output dimensions from 1 px up to 12,000 px per side (browser memory may limit very large outputs on older phones).

Aspect ratio: resizing changes dimensions, not the picture's shape — with the lock on, entering a width recalculates the matching height. If a form demands exact dimensions that don't match your image's shape (for example, a square photo from a wide shot), unlock the lock only if slight stretching is acceptable; cropping first usually looks better, and a cropper is on our planned tools list.

Enlarging: you can resize up, but software cannot invent detail that was never captured. An enlarged image will look softer than a photo originally taken at that size.

A practical example

A 4000×3000 phone photo resized to 1200×900 for a blog post keeps the same shape (both are 4:3), drops from several megabytes to a few hundred kilobytes, and loads noticeably faster — while still looking sharp on screens, since most blog layouts never display it wider than about 800–1200 px anyway.

Supported formats & honest limitations

Conversions: JPG ↔ PNG ↔ WebP in any direction, for files up to 25 MB. Dimensions are never changed by conversion — only the format (and, for JPG/WebP, the quality).

Transparency: PNG and WebP can store transparent pixels; JPG cannot. Converting a transparent image to JPG fills the transparent areas with white. If you need a different background colour, that requires an editor — free options include Photopea (browser-based) or GIMP (desktop).

Quality: converting to JPG or WebP is lossy, and converting from JPG cannot restore detail the JPG already discarded — a JPG converted to PNG is the same picture in a bigger file, not a higher-quality one. Animated images (animated WebP/GIF) are not supported: only the first frame would be kept, so this tool declines GIF input rather than silently breaking your animation.

Browser support: conversion uses your browser's own image engine. WebP output works in all modern browsers (Chrome, Edge, Firefox, Safari 14+). If a very old browser can't encode WebP, the tool will tell you instead of producing a broken file.

A practical example

A designer sends a logo as a 180 KB WebP, but a client's email signature tool only accepts JPG. Converting here at 90% quality produces a JPG of roughly 60–90 KB on a white background — ready to attach. If the logo relied on transparency over a coloured background, PNG would be the safer target instead, and the tool keeps the transparency intact.

Supported formats & honest limitations

Works with: JPG/JPEG, PNG, and WebP input up to 25 MB; output as JPG, PNG, or WebP at the selection's exact pixel size.

Cropping is permanent in the output: the pixels outside the selection are not in the downloaded file. That is the point of cropping — but it is also why you should keep your original. This tool never modifies your original file.

Small crops get soft when enlarged: cropping a 400×300 corner out of a photo gives you a 400×300 image. If you then print or display it large, it will look soft — cropping can't add detail that isn't inside the selection.

A practical example

A 4000×3000 product photo needs to become a square marketplace image. Set the ratio to 1:1, drag the selection over the product, and crop: you get a clean 3000×3000 square with the subject centred the way you chose — instead of whatever automatic centre-crop the marketplace would have applied.

Supported formats & honest limitations

Works with: JPG/JPEG, PNG, and WebP images, up to 25 MB each. Any number of images can be combined; each becomes one page, in list order.

Images are embedded as JPEG: inside the PDF, each image is stored as a high-quality JPEG. Transparent PNGs are placed on a white background (PDF pages are not transparent), and fine detail is preserved by a generous quality setting — but if you need a bit-perfect archival copy of a PNG, a PDF is the wrong container.

Size cap for sanity: photos larger than 2500 px on the longest side are scaled down inside the PDF. That is far beyond what A4/Letter printing needs (about 300 DPI at full page), and it keeps PDFs emailable instead of enormous.

No OCR: the PDF contains your images as pictures. Text in them is not searchable or selectable, because no text-recognition is performed.

A practical example

You photographed six pages of a signed agreement with your phone. Add all six, leave page size on A4 and orientation on Auto, and create the PDF: one tidy 6-page document, each photo upright and centred, ready to email or upload to the portal that asked for “a single PDF”.

What the numbers mean & honest limitations

Pixel dimensions are the image's true size — everything else derives from them. Aspect ratio is the width-to-height relationship, simplified like a fraction: a 1920×1080 image is 16:9. Megapixels are simply width × height in millions — a rough detail measure, not a quality score.

Transparency is reported from a quick scan of the image's pixels. JPGs never have transparency; PNGs and WebPs may. The scan samples the image rather than reading an embedded flag, which keeps it fast on large photos.

Not shown: print DPI. Some files carry a “dots per inch” metadata tag, but it is only a printing hint the file declares about itself — it doesn't change the pixels, and browsers don't expose it reliably. What matters for print is pixel dimensions divided by the printed size; 3000 px across a 10-inch print is 300 DPI.

A practical example

A marketplace rejects your upload saying “image too small”. The checker shows 800×800 px — below the 2000 px the marketplace wants for zoom. The fix isn't here (this tool only measures), but now you know the real problem: re-export or re-shoot at a larger size rather than guessing, and use the Resizer when you need exact output dimensions.

Supported formats & honest limitations

Input: JPG, PNG, or WebP up to 25 MB. PNG with transparency is the best source — the transparency is preserved in every generated icon. JPG and WebP sources are flattened as-is (convert a transparent PNG if your logo needs a see-through background).

Output: one favicon.ico (a standard ICO container holding real PNG images at 16/32/48 px — the format modern browsers expect) and individual PNG files at 16, 32, 180, 192, and 512 px.

What this doesn't do: it doesn't design a logo, simplify detailed artwork, or generate the full platform manifest some web apps use. If your source is a detailed photo, the 16 px icon will be unreadable no matter who generates it — start from a simplified version of your mark.

A practical example

A blogger's logo is a 1000×1000 PNG of a bold letter “B”. Generated here, the tab icon stays crisp because the shape is simple and the file is square. The 180 px Apple touch icon is what appears when a reader saves the blog to their iPhone home screen — a size most people forget, included in the set automatically.

Supported formats & honest limitations

Works with: JPG/JPEG, PNG, and WebP input up to 25 MB; output as JPG, PNG, or WebP at exactly the dimensions selected.

Fill mode crops: to fill a differently-shaped frame, the edges of your image are cut away — top/bottom for wide targets, sides for tall ones. The crop is always centred. If the subject sits at the edge of your photo, crop it deliberately with the Image Cropper first, then resize here.

Platforms change their specs: the presets reflect the sizes these platforms currently document and accept. Platforms occasionally retire sizes or crop differently in specific placements (a feed vs. a profile grid, for example) — if a platform's own help centre states a different number, type it in directly.

Upscaling warning: a small image stretched to a big preset (an 800 px photo into a 2560×1440 banner) will look soft. Start from the largest original you have.

A practical example

A wide 4000×2250 travel photo becomes three posts in under a minute: a 1080×1080 Instagram square (Fill — the horizon stays, the sides trim), a 1280×720 YouTube thumbnail, and a 1000×1500 Pinterest pin. One master photo, three platform-perfect files, zero surprise crops.

Supported formats & honest limitations

Works with: PNG and WebP input (both can carry transparency) up to 25 MB; output is always JPG at the same pixel dimensions. For JPG, GIF-free WebP, or other directions, use the general Format Converter.

The background is permanent in the JPG: once transparency is filled with a colour, the JPG has no way to “un-fill” it. That is a property of the JPG format itself, not this tool — keep your PNG if you might need the transparency again.

Edge fringing: anti-aliased edges (the soft semi-transparent pixels around a logo's outline) blend with your chosen colour. If the JPG will sit on a dark page, fill with that dark colour here; filling with white and placing on black leaves a pale halo.

A practical example

A product cut-out PNG on transparency has to go onto a marketplace that only accepts JPGs on white backgrounds. Converting here with the white background produces exactly what the marketplace would have created badly on its own — a clean white-background JPG, with the product's soft shadow blended naturally.

Honest limitations

One pixel, not an average: this tool reads the single pixel you tap. In photographs, neighbouring pixels vary (grain, shading, JPEG artifacts), so treat a photo colour as “that exact spot” — tapping a few points nearby gives you the practical range. Logos and flat graphics are exact.

Colour space: values are the sRGB colours as your browser decodes the image — the same numbers your design tool will show for the same file. A wide-gamut (Display P3) photo may contain colours sRGB numbers can't fully express; the HEX is still the closest standard value.

Transparency is reported, not hidden: if you tap a transparent or semi-transparent pixel (PNG/WebP), the opacity is shown honestly next to the colour. A fully transparent pixel's stored colour is meaningless, and the tool says so.

A practical example

A client sends their logo as a PNG and asks you to match its blue in a document. Open the PNG here, tap the blue, and copy #1B4F9C (or whatever your tap returns) straight into your document's theme — no eyedropper plugin, no guessing from a screenshot.

Supported formats & honest limitations

Works with: single-photo HEIC/HEIF files from iPhones and iPads, up to 25 MB. Output is JPG at the photo's original pixel dimensions.

How the decoding works: Safari (on Apple devices) can decode HEIC natively, and the tool uses that when available. Chrome, Edge, and Firefox cannot, so for them the page loads heic2any — a free, open-source decoder — from a public CDN, and the decoding still happens locally in your tab. If the decoder can't be loaded (for example, you're offline) the tool will tell you, rather than failing silently.

What doesn't survive conversion: “Live Photo” motion (only the still image converts), depth maps used for Portrait-mode re-editing, and HDR metadata. The picture itself — at full resolution — converts faithfully. If a HEIC holds a burst or several images, the first image is converted.

Size expectations: HEIC is roughly twice as efficient as JPG, so a converted JPG at high quality is commonly 1.5–2× the HEIC's file size. That's normal, not a fault — JPG trades storage for compatibility.

A practical example

A relative emails you iPhone photos for a family album you're assembling in older desktop software that rejects every .heic file. Convert each photo here — ten seconds each — and the whole set drops into the album as ordinary JPGs, still at full iPhone resolution for printing.

Supported formats & honest limitations

Works with: JPG/JPEG, PNG, and WebP input up to 25 MB. Output is a transparent PNG at the image's original dimensions (very large images may be processed at a reduced size by the service, in which case the result states its real dimensions).

This tool needs the internet and our service: AI background removal requires a segmentation model that doesn't fit in a browser tab. If the processing service is unavailable or hasn't been configured for this site yet, the tool will say so clearly instead of pretending to work — and your image simply isn't processed.

Edge quality varies: wispy hair, fur, transparent objects (glasses, veils, bottles), and subjects the same colour as their background can leave fringes or missing pieces. Professional product photography sometimes still needs a human touch-up; for typical portraits, products, and pets, results are usually clean.

A practical example

A seller photographs a handmade bag on a busy bedroom background. One pass here isolates the bag on transparency; placing that PNG on a clean white background (via the PNG to JPG Converter) gives a marketplace-ready listing photo — no studio, no manual tracing.

Honest limitations

It runs on a free shared queue: this generator is powered by a shared free service, so your prompt waits its turn alongside other people's. Most images arrive in 1–3 minutes; at busy times it can take longer, and if the queue is overloaded the tool will tell you plainly and ask you to try again — it never fakes a result. Leaving the tab and coming back cancels nothing on the server, but this page stops checking, so keep the tab open.

Three sizes only: square 512×512, landscape 576×384, and portrait 384×576. These are modest, web-friendly dimensions — great for posts, mock-ups, and drafts. For print or large displays, generate here and then enlarge carefully with the Image Resizer, knowing that upscaling softens detail.

AI images have AI flaws: extra fingers, garbled lettering, and odd small details are normal, especially in busy scenes. Prompts asking for readable text in the picture (“a sign that says…”) rarely come out clean — add your text afterwards with a design tool instead. Results also vary between runs: the same prompt twice gives two different images.

Use common sense with prompts: don't ask for images of real private people, and don't include personal information in your prompt text — it is sent to the generation service to do its work.

A practical example

A small shop owner needs a header image for a blog post about morning routines. The prompt “a wooden desk by a bright window, notebook, pen and a cup of tea, soft morning light, photorealistic” at landscape size returns a warm 576×384 header in about two minutes — free, with no stock-photo licence to check and no photo shoot to arrange. She then runs it through the Image Compressor so the page loads fast.

Honest limitations

Printed English text is the sweet spot: this tool is set up for English. Other languages, handwriting, and decorative fonts will give poor or empty results — the tool will show you exactly what it found rather than guessing.

Image quality decides accuracy: blur, glare, shadows, skewed angles, and tiny or low-contrast text all cause mistakes. If a result is messy, a straighter, brighter photo of the same page usually fixes more than any setting could. Cropping tightly around the text (with the Image Cropper) also helps.

Layout is simplified: columns, tables, and multi-column pages come out as plain reading-order text, not a recreated document. Check numbers digit by digit before relying on them — OCR can confuse 0/O, 1/l, and 5/S.

First use downloads the engine: on a slow connection the first extraction can take a while before reading even starts. Later uses in the same browser are much faster because the engine is cached.

Honest limitations

The built-in backgrounds are simple illustrations: they are drawn in your browser as a convenience starting point — they are not the famous meme photos (those belong to their photographers and studios). The real point of this tool is your own image: a pet, a reaction shot, a screenshot moment.

One caption style: the classic look — bold white capitals with a black outline — is the only style, on purpose. If you need fonts, stickers, or multiple text boxes, a full design app is the better tool.

Output size: memes are exported at the background's own size, capped at 1,200 px on the longest side, which is plenty for social posts and group chats. Very large photos are scaled down for the export, never up.

Mind the rights: only meme images you have the right to use. A funny caption doesn't transfer anyone's photo copyright or likeness rights to you.

Honest limitations

You mark the subject — no AI: phone portrait modes use depth sensors and AI segmentation to trace hair and edges. This tool deliberately does neither: it blurs everything outside the frame you draw. That's why it's free, private, and predictable — and why the sharp/blur boundary is a clean geometric edge rather than a traced outline.

Hard edge, not feathered depth: the transition between sharp and blurred is a single step. Subjects that roughly fill a rectangle or oval (a person, a product, a pet's face) look great; wispy hair over a detailed background will show the geometric edge, because the frame can't follow it.

Output size: the result is exported at your photo's own dimensions, capped at 3,000 px on the longest side for very large photos so phones don't run out of memory. The blur strength is scaled to match.

Guides in the research queue

The topics below are planned, in the owner's research order. They are listed as ideas to validate — checking real search demand and competing pages first — not as finished articles, and no traffic or ranking claims are made for them:

• How to Optimize Images for WordPress
• How to Prepare Product Images for an Online Store
• Best Image Compression Settings for Websites
• How to Make an Image Smaller for Email

Want one of these sooner? Tell us which — requests decide the order.