Getty & iStock controlled-vocabulary keywords in seconds. Not nights.

PixTagger is built by NoSystem Images, an exclusive Getty Images and iStock contributor since 2007, with 57,000+ accepted photos and 9,700+ video clips. We don't guess what stock contributors need — we live it.

Four steps: (1) Sign in and pick Photos or Videos mode in the dashboard. (2) Optionally type Shoot Notes (location, model, brief) — do this before you drop, so the context applies to the whole batch. (3) Drag a batch of files (or a whole folder) into the upload zone — tagging starts automatically the moment they land, so there's no separate "generate" button: we run vision AI per file (3 frames per video), filter Getty's controlled vocabulary, and build the title, description and keyword list. (4) Review the results in the table, edit any keyword inline if you want, then download the CSV or copy hashtags for Meta. For a mixed shoot, run two batches (one per mode). From drop to CSV: usually under 60 seconds.

Photos: JPG, JPEG, PNG, WEBP — up to 150 MB per file. Videos: MP4, MOV, M4V — these are the formats Getty and Adobe Stock accept (longer clips are sampled at start, middle, end). Each batch is one media type — pick Photos or Videos mode in the dashboard before dropping files.

No — tag them as-is. ProRes (including 422 HQ) works directly and is a great master format for Getty/iStock; large clips just take a little longer to process in the browser. Important: tag the exact file you'll submit. The file name, including its extension, in the exported CSV has to match the file you upload — Getty's DeepMeta and qHero match rows to files by full file name, so a converted "clip.mp4" would not attach to the "clip.mov" you deliver. Tagging your actual .mov keeps the names lined up automatically. One caveat if you want camera-movement keywords: measuring movement needs the browser to decode the clip, and Chrome and Firefox skip ProRes, HEVC and cinema-grade H.264 — tag those in Safari, or leave the box unticked and name the move in your shoot notes instead. Standard MP4 / H.264 is measured in any browser.

Yes. Each drop is capped at 100 photos or 30 videos — videos are lower because we decode three frames from every clip right in your browser before anything is sent, which is heavier on the tab than a photo resize. But that cap is per drop, not per export: keep dropping more batches and they stack in the same review screen and download as a single combined CSV. So if you have 200 clips going to Getty in one upload, drop them in batches and export one CSV at the end. A photo is 1 credit and a video clip is 2 (each clip runs a heavier multi-frame analysis) — 2 and 3 respectively when you tick Varied context — and re-dropping a file that's already in the export is skipped so you're never charged twice.

They would, and that's what Varied context is for. Tick "These are similar shots from the same set" before you drop a single shoot, and each file is told which angles its siblings already took so it leads with a different one — five frames of the same meeting can rank for five searches instead of splitting one. It never invents anything to look different: a frame that shows something else (a detail, an empty room, no people) is described on its own terms, keywords vary only in order and emphasis, and on video the camera-movement terms are left alone because those are facts. It costs 2 credits per photo and 3 per clip, takes a little longer, and only makes sense on a single shoot — on a mixed batch, leave it off.

That is what the Refine tool is for. Paste a keyword list, or upload a bulk CSV, and we map every term onto Getty's controlled vocabulary and top the list up with in-vocabulary terms your asset supports — so more of your keywords land as valid matches and more of them survive qHero refinement. No image is uploaded and none is needed: it works purely on the words. 1 credit per asset. It is the fastest way to bring an existing back catalogue up to standard without re-tagging it from scratch.

Yes, and they are treated differently from photographs on purpose. Tick "Illustrations / vector art" before you drop the batch, and two things change. We tag the concepts your artwork represents — Innovation, Teamwork, Growth — because that is what illustration buyers search for, rather than only the literal shapes on the canvas. And we switch on a separate vector vocabulary of terms the illustration trade actually uses (flat, isolated, pictogram, outline, stroke, icon set, UI), which Getty's photography-derived list simply never contained; measured against 7,577 live Getty illustrations it recognises about 21 more keywords on a typical file. Vector batches also export with .eps file names so they import into qHero without errors.

Yes — type them in whatever language you think in. Spanish, Russian, Chinese, German, Serbian and 40+ more all work, and you still get clean English titles, descriptions and keywords built for every marketplace. There is no need to translate your own notes into English first.

Yes, and it is free — tick "Use location (GPS)" and we read the coordinates your camera or phone saved in each file and tag the country, region and nearby place. It is worth having on any travel or landscape shoot, where the location is half of what a buyer searches for. Two things worth knowing: the coordinates are read on your own device before upload and are used for nothing except naming the place, and files without GPS are simply tagged as usual rather than guessed at.

From the photo's own EXIF capture time — the moment the shutter fired — read in your browser before the upload strips metadata. We deliberately do not fall back to the file's timestamp, because that is when the file was last written: any photo you have edited, exported or re-saved would carry a date months after the shoot, and agencies check the created date against the date on your model release and reject the file when the two disagree. If a photo has no EXIF date, we leave the column blank rather than write a wrong one. Video keeps the file timestamp, since clips do not carry the same tag.

Before — there is a "Tagging for" choice at the top of the dashboard, and it changes the output rather than just the file you download. Getty and Adobe want different things: Getty accepts only its own controlled vocabulary and up to 50 keywords, Adobe takes up to 49 free-text keywords and weights the first ten most heavily. Picking up front means the keyword list is built for that marketplace instead of being trimmed to fit it afterwards. You can still export the other formats from the same batch.

Tick "AI-generated" and we read the generation prompt saved inside the file and use it as context, and prime the tagger for surreal or hybrid subjects — an animal in a human role, an impossible landscape — which otherwise get described too literally. It applies to Adobe Stock, Shutterstock, Dreamstime and Pond5. Getty and iStock do not accept AI-generated content at all, so that combination is not something we can help with.

No. If a file cannot be processed, the credit is returned to your balance automatically and the row is marked with what went wrong, so a corrupt file or a dropped connection never quietly costs you anything. Dropping a file that is already in the current export is skipped rather than charged twice, so you can re-drop a folder without watching your balance.

Yes. Every file shows up in a results table with its generated title, description and keyword list. Click any keyword to remove it; type to add your own. Changes apply only to the export you download — your account history keeps the original AI output too, so you can compare or re-run.

Adobe CSV columns: Filename, Title, Keywords (max 49), Category, Releases. Getty CSV columns: file name, created date, title, description, country, brief code, keywords (max 50, all from Getty's official controlled vocabulary). Both formats are import-ready — drop them straight into Adobe Stock Contributor or Getty ESP without reformatting.

Direct CSV exports for Getty Images / iStock (ESP, DeepMeta and qHero formats), Adobe Stock, Shutterstock (with categories mapped to their official list), Pond5 and Dreamstime — each built to that marketplace's own column layout and keyword limits so the import works on the first try. For agencies that read metadata embedded in the file itself — Alamy, Depositphotos, 123RF — download XMP sidecars and let Lightroom or Bridge write the metadata into your exported files. Meta hashtags cover Instagram, Facebook and X.

An XMP sidecar is a small .xmp text file with the same name as your photo that carries its title, description and keywords in the industry-standard format Lightroom Classic, Bridge and Photo Mechanic read. Put the sidecar next to your original, read the metadata into your catalog, and every JPEG you export carries it embedded — which agencies like Alamy, Depositphotos and 123RF then pick up automatically on upload. It's the archive-first workflow: your catalog stays the single source of truth, and your originals never leave your computer — we generate sidecars purely from the metadata, without ever seeing your full-resolution files.

PixTagger isn't a single "describe this image" call. We run a layered pipeline tuned specifically for stock marketplaces: a vision pass writes a buyer-focused title, a two-sentence commercial description, and a keyword list ordered strongest-first (the leading keywords carry the most search weight on Adobe Stock and Getty). For video, camera movement is not guessed from the frames at all: tick "Detect camera movement" and we measure how the picture actually moved across the whole clip, in your browser, before anything uploads — so a slow gimbal drift is caught and a locked-off tripod shot stays correctly silent instead of collecting a move it never had. A second pass then maps your keywords onto Getty's official controlled vocabulary and checks each term against what's actually in the frame, so your Getty CSV passes validation without invented tags. We build on proven OpenAI vision models, but the value is in this stock-specific workflow — not the raw model.

For every video file, PixTagger extracts three representative frames (start, middle, end), runs vision AI on each, and merges the keywords. This catches scene changes and story arcs that single-frame tools miss. Camera movement is a separate, optional step — it is measured rather than read off the frames, which is why it is its own tick box. A clip costs 2 credits regardless of length. Three sample points describe one continuous action completely — but some clips have genuinely complex action: a long take that moves through a scene, an activity that turns into another, a reveal at the end, subjects entering in turn. For those batches, tick "This batch has complex action" before dropping and we sample five points per clip instead of three, for 1 extra credit per clip. More frames means more visual evidence, so the merged keywords describe the whole sequence rather than the parts three frames happened to land on. On a short clip with one continuous action it adds nothing — leave it off.

Yes — but in two separate batches. Pick Photos mode in the dashboard, drop your photos, generate the Photos CSV. Then switch to Videos mode and do the same for the videos. This matches how Getty and Adobe Stock actually accept submissions (separate uploads for photos and videos), so you'll be uploading them separately on the receiving end too. Your Shoot Notes can stay the same across both batches.

Yes. Our Getty CSV export uses Getty's official whitelist of approved keywords, so your submissions don't get rejected for invalid terms.

No. Files are processed and deleted within minutes. We never sell, share, or train on your images. Your work stays yours.

Every new account starts with 15 free credits and no card — enough to tag 15 photos or 7 clips and judge the output on your own work rather than ours. There is no subscription to cancel afterwards: PixTagger sells one-time credit packs, so if you never buy one, nothing happens.

Yes. Your account page has a personal invite link. When somebody signs up through it and tags their first 10 files, we add 50 credits to your balance and 25 to theirs, on top of the free credits every account starts with. If they later buy a credit pack, you get another 20% of it in credits. Nothing is paid the moment somebody registers — the credits arrive once they are genuinely using the tool, which is exactly why they are worth having.

No. PixTagger sells one-time credit packs, not subscriptions — there's nothing to cancel and no monthly billing. A photo costs 1 credit and a video 2 (2 and 3 with Varied context; a clip read at 5 frames instead of 3 adds 1 more), and credits stay on your account until you use them.