Bulk resize and rename photos online, in your browser

Published 16 September 2026 · by Tekiba

Resizing: one number for the whole batch

A batch of camera photos usually mixes landscape and portrait frames, and setting a width across that mixture leaves the portrait ones taller than the rest. A longest edge treats them alike: 1600px means the long side becomes 1600px, whichever side that is, so every photo fits the same square. Width, height and percentage are there too.

The aspect ratio is always kept — both edges scale by one factor and round to a whole pixel, so nothing is stretched or cropped. And nothing is ever enlarged: ask for 4000px from a 1000px photo and it stays at 1000px rather than inventing detail.

The free version holds the result to 2048px on the long edge whatever mode you picked, Keep the original size included: a 4032×3024 photo becomes 2048×1536. If 2048px is the size you are after for a social feed, resize photos to 2048px for Facebook works through that job, and through what we could and could not verify about the number.

Why phone photos come out sideways in other tools

Turn a phone on its side and it does not turn the pixels. It writes the frame as the sensor read it and adds a tag — the EXIF Orientation field, 1 to 8 — saying which quarter turn and mirror to apply when the photo is shown. Your viewer reads that tag; a batch tool that ignores it hands you a hundred photos on their side.

Browsers are on the right side of this. Every engine we measured — Chromium, Firefox and WebKit — applies the orientation tag itself when decoding a JPEG, whether or not the page asks: imageOrientation, the option meant to turn that behaviour off, does not stop Chromium rotating the image. We measured that rather than relying on documentation that still describes an older meaning of the option. SignetDrop asks for the tag and then checks it was applied — a JPEG's frame header gives the size as stored, so a non-square photo that comes back at exactly those dimensions while the tag calls for a quarter turn was not turned, and the canvas turns it instead.

Two consequences:

• The long edge is the one you can see. A portrait photo stored as 4032×3024 with a quarter-turn tag counts as 3024×4032, so a longest edge of 1600 gives 1200×1600, not a sideways 1600×1200.
• The output is upright, tag or no tag. By default the whole EXIF block is gone: each photo is drawn onto a canvas and re-encoded. With the paid option to carry EXIF across, the orientation field in the copy is forced back to 1 — the pixels are already upright, and a surviving tag would turn the photo twice. Remove location data from photos before selling online covers the rest of what that block holds, and what is left in the output.

Three renaming patterns

Renaming happens on the way into the ZIP. Each example starts from IMG_4021.jpg in a batch of ten, saved as JPEG.

Numbering widens itself to fit the highest number: ten photos from 1 give two digits (kyoto-01 … kyoto-10) and eight give one, so names sort correctly without counting the batch first. A suffix keeps names recognisable where web-sized copies sit beside the originals. Templates take {name} {n} {n:3} {date} {time} {w} {h} {ext} — the original name, the position in the batch, when it ran, the output size. Anything else survives into the file name, tidied only as far as a ZIP requires, so a misspelling shows up instead of quietly disappearing.

Two rules apply to every pattern. The extension always comes from the output format, so saving the example above as WebP gives IMG_4021-web.webp rather than a file that will not open. And names are made safe for a ZIP: separators, control characters, Windows reserved names such as CON, leading dots, trailing spaces. Where several photos land on one name the first keeps it and the rest become -2, -3, compared ignoring case as extractors do: one photo silently overwriting another is the worst thing a batch tool can do.

The same pass can watermark and convert

Resizing, renaming, converting and watermarking all happen in one pass, so it costs one decode and one encode rather than four trips through separate tools. Convert to JPEG, PNG or WebP, with a quality slider for the lossy two, and add a text mark or a PNG logo in one of nine positions. The companion article covers watermarking: watermark multiple photos without uploading them, including how to check in your browser's developer tools that your photos never left your machine.

Where the free version stops

The free version does ten photos at a time, up to 2048px on the longest edge, with resizing, format conversion, numbering, suffixes and a watermark. No time limit, no account, and it never writes our name onto your photos — the only mark that comes out is yours.

The paid version is $29 once, not a subscription. It lifts the limits we impose — any number of photos, at any size, subject to what your browser and its memory can manage — and adds name templates, saved presets, a text mark and a logo together, padding and 1:1 / 4:5 / 16:9 canvases, and EXIF carried across for JPEG output. The full split is on the pricing panel.

Questions

Open the tool, read about watermarking a batch or whether eBay and Etsy accept a marked listing photo or squaring a batch without cropping it, fitting square photos to a 3:4 frame or whether resizing removes EXIF, or see what the page loads and counts.