PhotoTag.ai and Rastock AI both generate titles, descriptions and keywords for stock files, and they part company immediately after that. PhotoTag.ai writes the metadata into your files and hands them back for you to upload yourself. Rastock AI writes the metadata and then delivers the files to 10+ agencies over FTP. Which one fits depends on whether you already have a delivery workflow you are happy with.
That is the comparison in one paragraph, but the paragraph hides a real decision. A metadata generator and a distribution pipeline solve different halves of the same job, and buying the wrong half means you still do the tedious part by hand. This article is written by the team behind one of the two tools, so weigh the framing accordingly. The facts below come from each vendor's own documentation, and where a figure could not be verified from a primary source it is described rather than invented.
What each tool actually does
PhotoTag.ai is a metadata generator. You upload files to its web app, it produces a title, a description and a keyword set for each one, and you download the files with IPTC, XMP and EXIF data already embedded, or export a CSV in an agency's format. It does not submit anything to an agency. Its own how-to articles tell you to open an FTP client and move the files across yourself.
Rastock AI is a metadata generator with a delivery layer attached. The generation step is comparable. The difference is what happens next: files go out over FTP or SFTP to 10+ agencies, with a separate metadata profile per agency, submission status tracking, and rollback if a batch goes out wrong.
If you only ever upload to one agency and you like your current habit, that delivery layer is worth nothing to you. If you upload to six, it is most of the value. Almost everything else in this comparison follows from that one question.
The tools worth comparing
Anyone evaluating PhotoTag.ai should look at the same handful of tools, because they fall into three genuinely different categories and the category matters more than the feature list. Pure metadata generators stop at the file. Metadata-plus-delivery tools carry it to the agency. Revenue-share intermediaries sell on your behalf and keep a percentage.
PhotoTag.ai — metadata generation, no delivery
Best for contributors who want good metadata written into their files and already have uploading sorted, or who want to call metadata generation from their own code. The pricing model is pay-once credits rather than a subscription: credits do not expire, and new accounts get ten free upload credits without a card. One credit covers one photo or one video, and a video costs a single credit regardless of its length. The vendor's price table is rendered client-side and could not be read directly for this article, so no dollar figures are quoted here — check the pricing page yourself rather than trusting a roundup.
What it does well: genuine video support with QuickTime metadata written natively, a REST API, a Lightroom Classic plug-in, output in more than fifteen languages, and broad file-type coverage including SVG and TIFF. Batch size scales with how many credits you have bought, running from 500 files up to 1,500 at the largest tier. Controls include a maximum keyword count, output language, maximum title and description length, and a custom-context field for feeding the model your own information.
The main limitation is the one that defines the product: no FTP and no agency delivery of any kind. What its documentation calls agency integrations are CSV export templates for Shutterstock, Adobe Stock, Freepik, Dreamstime, Storyblocks and Motion Array — useful, but the submission is still yours to perform.
Rastock AI — metadata plus multi-agency delivery
Best for contributors sending the same files to several agencies who want each agency's rules applied automatically. Pricing is a subscription, with a 14-day free trial that does not require a card. Delivery runs over FTP or SFTP to 10+ agencies — Adobe Stock, Shutterstock, Magnific, Alamy, Depositphotos, Pond5, Canva, 123RF, VectorStock and Vecteezy — and any other agency that accepts FTP can be added manually.
What it does well: per-agency keyword caps and banned or mandatory keyword lists applied before submission rather than discovered at review, rollback when a batch goes out wrong, and submission status per file per destination. There is no revenue share and no ownership claim on your files, and IPTC and CSV export stay open, so leaving does not cost you your metadata.
The main limitation mirrors PhotoTag.ai's. If you upload to a single agency, most of what you are paying for sits idle, and a pure metadata generator will serve you better and probably cheaper. A subscription is also the wrong shape for genuinely seasonal output.
Xpiks — the curation-first desktop option
Best for contributors who want to curate AI output rather than accept it, and who would rather own a tool than rent one. Its pricing page lists a free Basic tier capped at fifteen files per session, a Pro licence at €49 one-time covering a year of unlimited files, and Pro+ at €99 a year adding reverse image search, all plugins and cloud upload. The Pro+ trial runs fourteen days without a card.
It has the best keyword-craft controls in this group by some distance: a persistent blocklist of words the AI may never apply, concept hierarchy expansion, a keyword count limit, and a quick-insert mode for accepting suggestions selectively instead of all at once. It uploads too, over FTP, FTPS and SFTP to any agency. AI keywording runs on credits — one per file for keywords alone, five per file with title and description — with a monthly allowance included at Pro+.
One correction worth making, because it appears almost everywhere: Xpiks is widely described as open source. It was, historically. The current site carries an Intmaker OÜ copyright, ships under a proprietary licence agreement, and its GitHub link points to an issue tracker rather than source code. It remains an excellent tool. It is simply not open source today.
StockSubmitter — the widest agency coverage
Best for Windows contributors who want to reach more agencies than anything else here can. StockSubmitter sells a subscription metered by submission, counted separately for each agency. Uploading files is free and unlimited; only sending a file for review consumes quota. The free tier is genuinely usable rather than a teaser — 33 submissions a month at each of Shutterstock, Adobe Stock and most others, and unlimited submissions at iStock ESP, Dreamstime and YayImages.
Around forty agencies are supported, it has been running for over a decade, and agency passwords are stored locally on your own machine rather than in a cloud service, which matters to contributors who will not hand credentials to a hosted tool. The limitation is blunt: Windows only, 64-bit. Mac and Linux users are pointed at its browser-based sibling instead.
Microstock+ — the same coverage, in a browser
Best for Mac and Linux contributors, and for anyone on a slow connection. Microstock+ comes from the same company as StockSubmitter but runs in a browser, and its uploads execute from Microstock+ servers rather than your machine — so delivery speed does not depend on your own bandwidth. For anyone pushing 4K footage from a domestic connection, that is a bigger deal than any feature on a comparison chart.
Onboarding is unusually generous: free unlimited submissions until you have earned $50 from the agencies. Its AI keywording is billed on actual model token usage rather than credit packs, which makes it the most transparent billing in this comparison and refreshingly honest about how cheap inference has become.
BlackBox — footage, and paying your collaborators
A different category altogether. Membership is free and BlackBox takes 15% of net sales, charged only when something actually sells. It handles video only — no photos — and distributes to five agencies: Shutterstock, Adobe Stock, Pond5, Storyblocks and VimeoStock. Creators keep ownership of their work.
Its distinguishing feature is Shares: fractional ownership of a clip split across everyone who worked on it — the camera operator, the editor, the keyworder, the models — with the downstream payment splits handled automatically and in perpetuity. No metadata tool in this comparison does anything remotely similar, and for a small production team it solves an accounting problem that otherwise falls on whoever holds the account.
Wirestock — check the date on anything you read
Wirestock appears in nearly every best-tools list as a multi-agency distributor. In an announcement on its own blog in April 2026, it said it had stopped distributing content to third-party stock marketplaces, was retiring the Portfolio and Marketplace, and had discontinued Premium plans with refunds for the unused portion. It now positions itself around AI training data and dataset licensing.
If you are comparing tools using anything written before mid-2026 — including our own earlier Wirestock comparison — that section describes a service that no longer works the way it is described. It is a useful reminder that this category changes faster than the articles about it.
When to use PhotoTag.ai and when to use Rastock AI
Both generate metadata well enough that quality is rarely the deciding factor. The choice is about the shape of the rest of your workflow.
Choose PhotoTag.ai if you upload to one or two agencies and are comfortable with your current FTP habit; if you want metadata generation available over an API to call from your own scripts; if you work inside Lightroom Classic and want the step to live there; if your output is irregular enough that a subscription clock would waste money in quiet months; or if you need output in a language other than English.
Choose Rastock AI if you upload to four or more agencies; if you have been caught by an agency-specific rule such as a keyword cap or a banned term and want it enforced before submission rather than discovered at review; if you need to know which files landed where without opening six dashboards; or if you want a bad batch to be reversible.
Choose neither if your real problem is distribution breadth rather than metadata quality — StockSubmitter and Microstock+ reach far more agencies than either of these. And if you shoot footage with a crew, look at BlackBox before any of them, because none of the others will pay your editor.
Where PhotoTag.ai is genuinely stronger
Credits that never expire is the fairer model for anyone whose output is seasonal. A wedding photographer working through a summer's shoots in October is punished by a monthly subscription and not at all by a credit balance. A subscription is the wrong shape for that pattern, and it would be dishonest to pretend otherwise.
The API and the Lightroom plug-in matter more than they look. They mean PhotoTag.ai slots into a pipeline somebody else designed, rather than asking you to adopt its own. Any tool that assumes it is the centre of your workflow is a worse fit for a contributor who already has one that works. And one credit per video regardless of duration is a genuinely generous billing unit for anyone submitting long clips.
Where Rastock AI is stronger
Per-agency policy enforcement is the first. Shutterstock caps keywords at fifty and bans trademarks in commercial metadata; Adobe Stock caps at forty-nine and weights the leading keywords. One metadata set pushed to both will be wrong for at least one of them. Applying each destination's rules before submission is the difference between a clean review queue and a slow drip of rejections you have to fix by hand.
Delivery and status is the second. Generating metadata has become the easy part for everyone in this list. Getting thousands of files to ten destinations and knowing what happened to each one is the part that still consumes evenings. The third is reversibility: a metadata generator's mistake lives in your own files, where you can fix it, while a distribution pipeline's mistake lives on ten agencies' servers, which is exactly why rollback stops being a luxury once delivery enters the picture.
What neither tool does
Neither sells your work or takes a cut of it, and neither claims ownership of your files — which sets both apart from BlackBox and from Wirestock's former model. If that distinction is what is driving your search, the revenue-share question is worth reading about on its own.
Neither will rescue a portfolio that is not selling because of what is in the frame. Metadata decides whether a buyer can find your file. It does not decide whether they want it. Tools in this category are routinely oversold on that point, and a contributor whose downloads are flat because the work is generic will not be saved by better keywords. Neither guarantees acceptance either — every agency reviews content as well as metadata, and no tool has any influence over the first of those.
Testing both properly
Both are cheap to try honestly. PhotoTag.ai gives ten free credits with no card. Rastock AI runs a 14-day free trial, also with no card. Put the same forty files through each and compare more than the keyword quality — time the whole job end to end, uploading included, because that is where the two tools actually differ. The full feature list and current plans are both public.
We have written the same comparison against Xpiks and StockSubmitter, and a broader survey of the keyword generation field as a whole covers the tools that did not fit here.
The honest summary is that PhotoTag.ai is the better metadata generator to drop into somebody else's workflow, and Rastock AI is the better fit if you want the workflow itself. Those are different products that happen to share a first step, and the contributors who are happiest with either one are the ones who worked out which half of the job was actually costing them time.
Frequently asked questions
Does PhotoTag.ai upload to stock agencies?
No. PhotoTag.ai generates metadata and either embeds it into your files or exports a CSV in an agency's format, but it does not submit anything. Its own guides instruct you to open a separate FTP client and transfer the files yourself. What its documentation calls agency integrations are export templates rather than delivery, and that is the single biggest difference between it and tools that include FTP.
Is PhotoTag.ai a subscription?
No, it uses pay-once credits. Credits do not expire, and new accounts receive ten free upload credits without a card. One credit covers a title, description and keyword set for one photo or one video, and a video costs a single credit regardless of its length. That model suits irregular output better than a monthly subscription, because quiet months cost nothing at all.
What is the main difference between Rastock AI and PhotoTag.ai?
Delivery. Both generate titles, descriptions and keywords of comparable quality. PhotoTag.ai stops once the metadata is in your files. Rastock AI continues, applying each agency's rules such as keyword caps and banned or required terms, then delivering over FTP to 10+ agencies with per-file status tracking and rollback. If you upload to one agency that layer is wasted; if you upload to six, it is most of the value.
Which tool is better for video metadata?
PhotoTag.ai has strong video handling for a pure metadata tool: MP4, MOV and AVI, with QuickTime metadata written natively, and one credit per video regardless of duration. For footage specifically it is also worth looking at BlackBox, which handles video only and distributes to Shutterstock, Adobe Stock, Pond5, Storyblocks and VimeoStock in exchange for 15% of net sales.
Is Xpiks still open source?
No. Xpiks was open source historically, which is why the claim survives on aggregator sites, but the current site carries an Intmaker OU copyright, ships under a proprietary licence agreement, and its GitHub link points to an issue tracker rather than source code. It remains an excellent tool with the best keyword curation controls in this category. It is simply not open source today.
Does Wirestock still distribute to stock agencies?
No. In an announcement on its own blog in April 2026, Wirestock said it had stopped distributing content to third-party stock marketplaces, was retiring its Portfolio and Marketplace, and had discontinued Premium plans with refunds for unused time. It now positions itself around AI training data and dataset licensing, so comparison articles written before mid-2026 describe a service that no longer operates that way.