A CUR file is a Windows cursor with its artwork locked inside a format nothing else can open (read our breakdown on what is a CUR file to learn more). Converting it to PNG is not a format fashion choice — it is the only way to see the artwork on a Mac, edit it in Photoshop or GIMP, or use it anywhere outside Windows.
This page does the extraction in your browser: your files are never uploaded, the alpha channel is preserved, and the hotspot coordinates are shown on every card so nothing about the original cursor is silently thrown away.
What makes a CUR file different from an ICO file
CUR and ICO share one binary container, and that shared ancestry is why converters treat them as cousins. Every file starts with a six-byte ICONDIR header: two reserved bytes, a type field, and an image count.
ICO files set the type to 1; CUR files set it to 2. After the header comes one 16-byte directory entry per embedded image, and here is where the formats diverge.
ICO entries use two of those bytes for color planes and bit count. CUR entries repurpose exactly those bytes to store the hotspot: the X coordinate where the planes value would sit, and the Y coordinate where the bit count would sit.
- ICONDIR — 6 bytes, reserved bytes, type field, image count
- ICO type 1, CUR type 2
- CUR hotspot — X in the planes bytes, Y in the bit-count bytes
The hotspot is the whole reason CUR exists as a separate format. It is the single pixel Windows treats as the click point — the tip of an arrow cursor at (0, 0), the center of a crosshair, the nib of a pen.
An icon never needs one, because nobody clicks with an icon. A cursor without a correct hotspot feels broken: the pointer appears to click a few pixels away from where the user aimed, which is exactly the complaint that fills cursor forums.
When a CUR holds several sizes — 16×16, 32×32, 48×48 are the classic set — each size can carry its own hotspot, and a well-built cursor scales the hotspot with the artwork rather than copying one coordinate across all sizes.
Takeaway: a hotspot is not decoration. It is the click point itself — get it wrong, and the cursor clicks where you did not aim.
Why convert CUR to PNG: editing the artwork in Photoshop and GIMP
The most common reason to convert is editing, and the workflow is the same whether you are fixing a theme-pack cursor or designing one from scratch. Photoshop and GIMP cannot open .cur files at all, so the CUR has to become a PNG before any real work happens.
Extract it here, open the PNG in your editor, and do the work that matters: recolor the artwork, clean up a noisy edge, sharpen a blurry tip, or resize the design for a high-DPI display. Export a fresh PNG when you are done.
The round trip is where most converters abandon you, because the PNG you just edited has no hotspot and no way to become a cursor on its own. That is what the companion PNG to CUR Converter is for: upload the edited PNG, set the hotspot — the tip for an arrow, the center for a crosshair — and get a working .cur back.
If you extracted with this tool, the original hotspot coordinates are printed on the card, so re-entering them is a copy-paste job instead of guesswork. The full artwork side of that workflow, including the hotspot rule that decides whether a cursor clicks where it points, is covered in the guide on how to make custom cursors in Windows 11.
Editing is not the only reason. A PNG is viewable on macOS and Linux, where .cur files are opaque binaries; it drops into documentation, slide decks, and web pages; and it archives cleanly, since PNG is a lossless, universally supported format while CUR is a Windows-only container.
- Viewable on macOS and Linux, where .cur files are opaque
- Drops into documentation, slide decks, and web pages
- Archives cleanly as a lossless, universally supported format
The alpha channel: how cursor transparency survives extraction
Cursors live on transparency. A cursor graphic is an arrow or a hand floating over whatever is beneath it — there is no white square around it, because the pixels outside the artwork are transparent.
That transparency is the alpha channel, and preserving it is the entire quality test for a CUR to PNG converter. Get it wrong and the extracted image arrives with a white or black box baked around the artwork, which is useless for editing and embarrassing in documentation.
Modern cursors make this easy: they store 32-bit RGBA images, where every pixel carries its own alpha value, and that alpha maps directly into the PNG. Older cursors are trickier.
They store 24-bit color or a paletted image plus a second bitmap called the AND mask — a 1-bit-per-pixel map where a set bit means transparent. The visible image (the XOR mask) and the transparency map (the AND mask) are stacked in one DIB, with the height field doubled to cover both.
This converter reconstructs alpha from both sources: real alpha bytes where they exist, AND-mask bits where they do not, including the common trap where a 32-bit cursor has all-zero alpha bytes and only the AND mask tells the truth. The card preview sits on a checkerboard precisely so you can verify transparency with your eyes before downloading — if the background shows through around the artwork, the alpha survived.
- 32-bit RGBA — real alpha bytes, mapped directly to PNG
- 24-bit or paletted + AND mask — transparency reconstructed from mask bits
- All-zero alpha trap — 32-bit cursor where only the AND mask tells the truth
This is also why PNG, not JPG, is the extraction target. JPG has no alpha channel at all: converting a cursor to JPG would flatten every transparent pixel to white (or black, depending on the encoder) and permanently destroy the shape of the artwork.
PNG is lossless and alpha-capable, so the extracted image is pixel-identical to the cursor's artwork, transparency included.
Takeaway: if the preview does not show the checkerboard through the background, the transparency did not survive — do not download that version.
What happens to the hotspot — and how to keep it
PNG has no field for a hotspot, so the coordinate cannot travel inside the PNG file, and any converter that claims otherwise is confused. What differs between converters is whether the hotspot is reported or silently discarded.
Convertio's CUR pages mention hotspot data in one line and then drop it; ImageToSTL never mentions it at all. This tool prints the hotspot X and Y on every FileCard next to the dimensions, because the number is the only bridge back to a working cursor.
Extract, edit, convert back with the same coordinates, and the rebuilt cursor clicks exactly where the original did. Lose the number and you are back to nudging a hotspot by trial and error.
Takeaway: if you lose the hotspot number, you lose the cursor. Copy the coordinates from the card before you edit.
Multiple sizes, animated cursors, and other edge cases
A CUR file is a container, not a single image: it can hold 16×16, 32×32, 48×48, and larger artwork side by side so Windows can pick the right size for the display. This converter extracts the largest embedded image, because that is the artwork worth editing — a 16-pixel thumbnail upscaled would only blur.
The card shows the extracted dimensions, so there is never ambiguity about which size you got.
- Classic sizes: 16×16, 32×32, 48×48, and larger
- Extracted: the largest embedded image only
- Shown on the card: extracted dimensions, plus hotspot X and Y
One correction to two ranking competitors: CUR files are static. Animated cursors are a different format with the .ani extension, and an .ani file fed to this converter will show an error rather than a PNG.
If you need animation, that is a different conversion entirely, and any page claiming to animate a .cur is selling a misunderstanding.
How to convert CUR to PNG
- Select your files. Drag .cur files onto the converter or press SELECT FILES. Each file appears as a card with a preview, its dimensions, and its hotspot coordinates.
- Press CONVERT. Each cursor image is extracted to PNG in your browser, alpha preserved, with no upload and no waiting on a server queue.
- Download. Press the green Download .PNG button on any card, or SAVE ALL to download every PNG individually. There is no ZIP file — you get plain PNGs, ready to edit or share.