Image Quarry — Extract, Convert & Snip

Extract every image on a page, convert between 40+ formats, compress to an exact file size, and snip the screen like a pro.

Image Quarry is a complete image toolkit that lives in your toolbar. Pull every picture off any web page, convert between dozens of formats, squeeze a file down to an exact size, and snip anything on your screen with a full annotation editor. Everything runs inside your browser — no uploads, no accounts, no servers. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ GRAB ONE IMAGE WITH A RIGHT-CLICK Right-click any image on any page and choose "Convert this image…". That exact image opens in the gallery, already ticked and pushed to the front of the grid, so it is the first thing you see instead of being buried among 150 others. Pick a format, hit download, done. The right-click menu also offers "Grab all images on this page" and "Snip a region of this page". GRAB EVERY IMAGE ON A PAGE Click the toolbar icon and press "Grab images from this page" (or press Alt+Shift+I). The scanner goes far beyond reading <img src>. It finds: - <img> tags, including currentSrc, every srcset candidate and the plain-src fallback - <picture> sources and their fallbacks - 17 different lazy-loading attributes (data-src, data-original, data-lazy-src, data-zoom-image and more) - CSS background-image, mask-image and shape-outside, both inline and from stylesheets - Inline <svg>, kept as vector - <canvas> contents, <video> posters, <object> and <embed> - Favicons, og:image and twitter:image social preview images - Open shadow roots and every same-origin iframe - Images that never appear in the page at all — the background worker watches network responses, so script-loaded textures and images fetched by XHR still show up DEEP SCAN scrolls the whole page first so lazy loaders actually fire, then merges both passes. RESCAN re-reads the page after it has changed. Then sort it out: filter by minimum width and height, by source kind, or by file name and URL. Sort by page order, largest first, smallest first, name or format. Select All, None, Invert or Large only, or just click any thumbnail to tick it. USE IT AS A STANDALONE CONVERTER No web page needed. Bring in your own files three ways: - Upload — the file picker lists every readable extension, so a .psd or a .cr2 isn't hidden behind "All files" - Drag and drop anywhere on the window - Paste with Ctrl+V, including a screenshot straight off the clipboard Your uploads sit alongside scanned images, are selected automatically, and survive a rescan of the page. CONVERT BETWEEN FORMATS Choose a target format, a quality level and a size, then download. PREVIEW ONE shows the source image beside the actual encoded result — the produced bytes are decoded back and painted, so a TIFF, a PSD or an EXR previews as a real picture and you see the true quality loss before committing. Thumbnails work the same way: formats no browser can display are decoded by the extension itself and drawn to a canvas. The built-in FORMAT SUPPORT page is generated at runtime by asking the codec engine what it can actually do, not from a hardcoded list — so you always know what is and isn't possible before you start. COMPRESS TO AN EXACT FILE SIZE Pick a preset (20 KB, 50 KB, 100 KB,)or type any custom KB or MB value. - Lossy targets (JPEG, WebP, AVIF) — binary-searches the quality setting so your dimensions survive - Lossless targets (PNG, TIFF, BMP and others) — dimensions are the only lever, so it scales down in coarse steps, then binary-searches between the last size that was too big and the first that fit, so a 60 KB budget lands near 57 KB instead of 40 KB - If even the lowest usable quality overshoots your target, it tells you plainly instead of quietly handing back something unusable RESIZE AND FINISH Set an output width and height, or a scale percentage. Choose Stretch, Contain or Cover, with keep-aspect-ratio on by default. Set a background colour for formats without transparency. Choose which sizes go inside a multi-size ICO or ICNS. Add a filename suffix such as "-web" to every file in the batch. Download converted files individually, or the whole selection as a single ZIP. SNIP YOUR SCREEN Press Alt+Shift+S, or use the popup. THE SCREEN FREEZES THE INSTANT YOU TRIGGER A SNIP, and you select on that frozen frame. That is the difference between capturing an open dropdown and capturing the page after it closed — the moment your mouse moves toward a toolbar, the page loses its hover state and the menu is gone. Capturing first keeps whatever was on screen: open menus, tooltips, hover styles, mid-animation states. Then take your time selecting it. Because the frame is a still image, the magnifier shows real pixels at 8x with a colour readout, and the page is locked against scrolling while you choose. Seven capture modes: - Rectangle — drag a box; hold Shift for a perfect square - Free-form — lasso any shape; everything outside the path becomes transparent - Window — hover to highlight an element or region, click to take it - Whole screen — the frozen frame exactly as it was - Full page — scrolls and stitches the entire scrollable page into one image - Full screen — any monitor, via the browser's screen picker - Application window — any other app's window, via the same picker THE ANNOTATION EDITOR Every snip opens in an editor with pen, highlighter, line, arrow, rectangle, ellipse (outlined or filled), text, numbered step badges, pixelate or blur redaction, a measuring tool showing pixel distance and angle, crop, undo, redo, clear and copy-to-clipboard. Hold Shift to constrain shapes to squares and lines to 22.5 degree increments. NOTHING IS BAKED INTO THE PICTURE UNTIL YOU SAVE. Every annotation stays a separate object you can pick up again with the select tool: - Click an annotation to select it — the toolbar switches to that item's own colour, thickness and font, so changing a control restyles it in place - Drag it anywhere; hold Shift to lock to one axis, or nudge with arrow keys (Shift + arrow moves ten pixels) - Drag a handle to resize — boxes, text and freehand strokes get eight handles; lines, arrows and the ruler get one at each end - Del removes it, Ctrl+D duplicates it, [ and ] send it behind or in front, Esc deselects - Every one of those is undoable with Ctrl+Z Text has a picker with 24 typefaces grouped into sans-serif, serif, monospace and display/handwriting, an explicit point size, and bold and italic toggles (Ctrl+B / Ctrl+I) — all stored per label, so two pieces of text can use two different fonts. Press Enter to place, Shift+Enter for a second line, Esc to discard. Double-click any text you already placed to edit the words again. Every typeface is one already installed on your machine, so nothing is downloaded and snips render the same offline. Saving from the editor offers the same format, quality, resize and target-file-size controls as the gallery. Selection outlines and handles are drawn on the canvas but stripped before anything is encoded, so they never reach a saved file or the clipboard. KEYBOARD SHORTCUTS Alt+Shift+I — grab images from the current page Alt+Shift+S — rectangular snip V select, P pen, H highlighter, A arrow, T text, B blur, C crop Ctrl+Z undo, Ctrl+C copy, Ctrl+S save, Ctrl+V paste an image in ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ FORMAT SUPPORT WRITES 48 FORMATS: PNG, JPEG, WebP, AVIF, GIF & many more. READS 61 FORMATS, including everything demosaiced from the sensor data, falling back to the embedded preview where a variant defeats the decoder. HONEST LIMITS, stated up front rather than discovered halfway through: - HEIC is read-only. The HEVC encoder is patent-encumbered and is left out of every WebAssembly build, so reading iPhone photos works but writing HEIC does not. - PDF and AI are write-only. Reading them would require an external program that cannot exist inside a browser. - RAW cannot be written in any build. A RAW file describes one specific camera sensor's readout, so it cannot be synthesised from a finished picture. The codec engine — ImageMagick compiled to WebAssembly — ships inside the extension. Nothing is downloaded at runtime, so conversion works completely offline. PRIVACY Everything happens locally in your browser. No image, no URL and no page content is ever sent anywhere. There are no analytics, no accounts and no server of any kind. The only network requests the extension makes are fetching the images you selected, from the same origin your browser already loaded them from. Site access is needed to read images from the page you are on. Downloads permission saves your converted files. Storage keeps your format and theme preferences. The webRequest permission is used in observe-only mode to note image URLs for the current tab — nothing is blocked, modified or retained beyond that tab's session. Light and dark themes included. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Created by Souvik Dey — Email Us -souvikdeymarketingpro@gmail.com © 2026 Image Quarry.