About OnlineCull
Last updated: October 15, 2026.
OnlineCull is a free RAW viewer and photo culler that runs entirely inside a browser tab. You point it at a folder, it shows the full-size preview embedded in every RAW within seconds, and you rate, pick and reject with the keyboard. It is built and maintained by Damien Yvert, an independent developer.
Why it reads your files instead of uploading them
The obvious way to build this would be to upload the shoot to a server, decode it there and send back thumbnails. That is a bad trade twice over. A 60 GB wedding takes hours to upload on a normal connection before the first frame appears — and a lot of photography is shot under an NDA, on a press pass, or in a hospital, where handing the files to a third party is not a decision the photographer gets to make casually.
So OnlineCull does the opposite. Your files are opened with the File System Access API, parsed in your browser, and never transmitted — there is no upload endpoint, so the site could not receive them even if it wanted to. You can confirm that in your own network inspector, which is the only form of privacy claim worth making.
It is fast for a reason that has nothing to do with our servers: every RAW file already contains a full-size JPEG that the camera rendered at the moment of capture. OnlineCull parses the container and pulls that image out. No demosaicing, no import step, nothing to wait for. It is the same trick that made Photo Mechanic famously quick, and it is why a browser tab can keep up with a desktop app here.
The honest consequence
Because you are looking at the camera’s own JPEG rendering, OnlineCull is not developing your RAW. That is enough to judge framing, expression and focus at 1:1 — which is what culling is — but it is not the sensor data, and no amount of browser cleverness would make it so. When you need to recover a highlight or fix white balance, that happens in Lightroom, Capture One, darktable or RawTherapee, after the cull.
Browsers also differ. Chrome and Edge can write XMP sidecars directly next to your originals; Safari and Firefox can open folders and cull perfectly well, but their exports arrive as downloads instead. We say so on the pages rather than discovering it together later.
What it deliberately does not do
Being clear about the edges is more useful than pretending there are none. OnlineCull does not edit photos, apply presets, crop or convert formats. It never modifies or deletes your originals — ratings go to sidecar files, never into the RAW. It does no IPTC captioning, no tethered shooting, no FTP delivery, and it does not read video. There is no API and no CLI.
It also has no AI auto-culling, and that one is a choice rather than a gap. Deciding which frame is the keeper is the part of the job that is actually yours; a tool that does it for you is solving a different problem. What OnlineCull tries to do is make the mechanical part — moving, zooming, marking — fast enough to disappear.
How the format pages are written
Each format page (CR2, CR3, NEF, ARW, RAF, DNG, ORF, RW2) is written against the parser that actually reads it, not from a spec sheet. Where a page says where a preview lives inside a container, that offset was read out of the implementation — which is why the ORF page explains that Olympus hides its previews in maker-specific corners of the container, and the CR2 page warns that the biggest JPEG in the file is lossless sensor data no browser can decode, so a viewer that just grabs the largest one shows you a grey rectangle. Those are the details that cost debugging time, so they are in the copy.
Contact
Corrections, bug reports and requests for a format we do not parse yet are all welcome. A report that names your camera model, browser and OS is worth ten that don't. Write to damienyvert.dev@gmail.com.
For exactly what the site does and does not collect, see the privacy policy. The short version is that your photos are never part of it.