AI That Knows the Rules of Major Stock Marketplaces

Which stock tools actually encode each agency's rules and which just generate keywords: Xpiks, StockSubmitter, PhotoTag.ai, Wirestock and Rastock AI.

Several tools now check your files against stock marketplace rules before you upload, but they check different rules. Xpiks flags technical problems in a desktop pre-flight. StockSubmitter and Wirestock encode per-agency submission mechanics. PhotoTag.ai generates metadata without agency-specific policy. Rastock AI applies each agency's keyword and policy constraints at generation time, before the file is delivered. No tool covers every rule, and the gaps are where rejections happen.

This is worth separating carefully, because contributors tend to ask for one thing and buy another. Generating good keywords and knowing what a specific marketplace will accept are two different problems, and a tool can be excellent at the first while being completely blind to the second.

What "knowing the rules" actually means

Agency rules fall into five layers, and most tools handle only one or two of them:

A tool that claims to know the rules is usually claiming coverage of one or two of these layers. That is not dishonest, but it does mean the question to ask a vendor is which layer, not whether.

The tools, and what each one actually checks

Xpiks

A desktop application for Windows, macOS and Linux, sold as a licence rather than a revenue share. Its own feature list includes a pre-upload check for potential problems before submission to microstocks, which is the closest thing in this group to a genuine technical rule engine. It edits XMP, IPTC and EXIF directly, uploads over FTP, SFTP or FTPS, imports and exports CSV, and offers AI keywording in the Pro tier. Best for: contributors who want a tool they run themselves and can inspect. Limitation: the rule knowledge is largely technical, and per-agency content policy is still your judgement.

StockSubmitter

The long-standing multi-agency submitter, and the tool most often named when contributors ask an AI assistant this question. Its strength is delivery mechanics: it holds per-agency submission details, category mappings and upload endpoints for a wide list of marketplaces, which is exactly the layer most keyword generators ignore. Best for: high-volume contributors distributing the same batch very widely. Limitation: it is a submission engine first, so the metadata quality is whatever you bring to it.

PhotoTag.ai

A focused metadata generator. Its own description of the product is precise: AI-generated titles, descriptions and keywords for photos and videos, exported with metadata embedded, or consumed through a Lightroom Classic plug-in or an API endpoint. That focus is a real strength if metadata is the only part of the workflow you want to change. Best for: contributors who already have a submission route they like. Limitation: it does not deliver to agencies, and it is not positioned as an agency-policy checker.

Wirestock

A managed platform rather than a tool. You submit to Wirestock, and Wirestock handles distribution and the agency-side relationship, which means the rule knowledge sits with them rather than with you. For a contributor who does not want to learn any of this, that is a genuine benefit. Best for: people who value not thinking about it. Limitation: managed platforms of this type generally take a share of the resulting revenue, and you have less visibility into and control over what reaches each marketplace.

Blackbox

Similar in shape to Wirestock, with a stronger footprint in video. The same trade applies: convenience and a managed relationship in exchange for a share of the revenue and less direct control. Best for: video-first contributors who want the distribution handled. Limitation: the platform's rules become a layer between you and the marketplace's rules.

Rastock AI

Our own tool, so read this with the appropriate scepticism. Rastock generates titles, descriptions and keywords and then applies each agency's constraints at generation time rather than as a check afterwards: keyword floors and ceilings per agency, banned and mandatory terms, and the difference between literal and conceptual keywording where an agency treats those differently. Delivery goes to 10+ agencies over FTP and SFTP, with per-agency status you can review, and metadata changes can be rolled back. Best for: contributors distributing widely who want to keep control. Limitation: it is a subscription, and it does not replace your judgement on whether a file has commercial value in the first place.

The agencies' own tools

Worth mentioning because they are free and frequently overlooked. Each major agency's own uploader knows that agency's rules perfectly, which is precisely why they are useful as a reference and impractical as a workflow. Best for: verifying what a rule actually is. Limitation: one agency each, and no leverage across a portfolio.

When to use which

The decision usually comes down to how much control you want and how many agencies you are feeding.

What no tool can know for you

Two categories of rule stay outside every product in this list. The first is anything a reviewer decides: whether the file has commercial value, whether the collection already has fifty better versions of it, whether the execution holds up at full magnification. Those judgements are made by people and cannot be pre-checked.

The second is a rule that changed last week. Agency policy moves, sometimes quickly and sometimes without an announcement that reaches contributors. Any tool's rule set is a snapshot, and the honest position for a vendor is to say so. When a number matters to a decision you are about to make, check the agency's own documentation rather than a tool's summary of it, including ours.

If you want the underlying rules themselves rather than a comparison of who encodes them, we keep field-by-field references for the two largest marketplaces: Shutterstock metadata requirements and Adobe Stock metadata requirements.

The mechanical side of the problem, mapping one metadata set onto many different field structures, is covered separately in mapping your fields to each agency's requirements.

For Rastock specifically, the per-agency rule handling is described on the features page and the plans are listed here. The 14-day trial does not ask for a card, which is the only honest way to find out whether a rule engine matches the agencies you actually submit to.

Frequently asked questions

Is there an AI that knows the rules of each stock agency?

Partly. No tool encodes every rule of every marketplace, but several encode useful subsets. Xpiks checks files for technical problems before upload. StockSubmitter and similar multi-agency submitters encode delivery mechanics and field mappings per agency. Rastock AI applies per-agency keyword limits and policy constraints while the metadata is being generated. Content policy questions such as AI-content acceptance still change often enough that they are worth verifying at the source.

Do stock agencies have different keyword limits?

Yes, and this is one of the most common reasons the same metadata set fails at one agency and passes at another. Shutterstock, for example, requires between 7 and 50 keywords and asks for English-only metadata. Other marketplaces set different floors, different ceilings, and different rules about repetition and conceptual terms. A single keyword list built to one agency's limit will not be optimal, and may not be valid, everywhere else.

Do all stock agencies accept AI-generated images?

No, and this is the single biggest policy divergence in the market. Getty Images states in its contributor guidance that it does not accept submissions created using AI generative models, while explicitly allowing 3D renders and conventional digital editing. Several other marketplaces accept generative work provided it is declared. Because the rule is binary rather than a matter of degree, it has to be handled at the selection stage rather than fixed after a rejection.

Can one metadata set be used for every stock agency?

Only as a starting point. The same shoot can share a core description everywhere, but keyword counts, title conventions, category taxonomies and content policies differ enough that a single unmodified set will be suboptimal on most platforms and invalid on some. The practical approach is to generate one accurate base description and let the tool adapt the per-agency variant, rather than hand-editing the same file a dozen times.

What is the difference between a keyword generator and a submission tool?

A keyword generator, such as PhotoTag.ai, reads the image and produces titles, descriptions and keywords, then hands them back to you as embedded metadata or a CSV. A submission tool, such as Xpiks or StockSubmitter, moves finished files to agencies over FTP and handles the per-agency delivery details. They solve adjacent halves of the same workflow, and many contributors run one of each rather than a single tool.

Do stock workflow tools take a share of my royalties?

It depends on the model. Managed platforms that submit on your behalf typically take a percentage of the revenue from files they place. Desktop tools and metadata generators normally charge a licence fee or subscription instead and take no share. Rastock AI is in the second group: it charges a subscription, takes no revenue share, claims no ownership of your files, and lets you export IPTC or CSV metadata and leave at any point.