What Is a GS1 Resolver?

Scan the QR code on a modern product and your phone opens a web page. But the code was printed months ago, possibly years before the page you are reading was written. Something in between decided where that scan should go today. For a GS1 Digital Link, that something is called a resolver.

This guide explains what a resolver is, how it chooses a destination, and where its job ends.

A resolver in one paragraph

A resolver is a web service that receives a GS1 Digital Link — a web address built from a product's GS1 identifiers — and connects the item it identifies to one or more online resources. Usually it answers with a redirect: "for this product, go here". The rules for how a GS1 resolver should behave are set out in the GS1-Conformant Resolver Standard, currently Release 1.2.1 (August 2026).

The key idea is separation. The packaging carries a stable identity. The resolver holds a set of links for that identity, and those links can be changed at any time without touching a single printed pack.

What a resolver receives

A GS1 Digital Link looks like an ordinary web address, but its path is structured. Here is one for a fictional perfume house, Maison Ardent:

https://id.maisonardent.example/01/09506000134352/10/LOT2610/21/A7F3K9

Reading it left to right:

• id.maisonardent.example is the domain the brand chose to print. Any domain is allowed.
• /01/09506000134352 is the GTIN, the number for this product type, written as 14 digits.
• /10/LOT2610 is the batch or lot.
• /21/A7F3K9 is the serial number of this individual bottle.

The resolver's first job is to read those identifiers and work out which item is being asked about. If you want the syntax in more depth, start with what GS1 Digital Link is.

What happens when someone scans

Take the same Maison Ardent bottle, carrying that address in a QR code.

1. A shopper scans. The phone's camera decodes the QR code and opens the address in a browser, like any link.
2. The request reaches the resolver. The resolver reads the GTIN, batch and serial from the path.
3. It looks up the links registered for that item. Maison Ardent has registered a product page, care instructions and a product registration form.
4. It picks one. The shopper did not ask for anything in particular, so the resolver uses the item's default link — here, the product page — and redirects the browser to it.
5. The page loads. The shopper never sees the resolver; they simply land on the right page.

The standard requires exactly one default link for each identified item. That rule matters: it means every ordinary scan has somewhere sensible to go. The resolver redirects to the default unless the request asks for something else.

There is one refinement. An optional default called gs1:defaultLinkMulti lets the resolver choose among several defaults based on language or other details the browser sends with its request. A shopper whose phone is set to French can be sent to the French product page without the brand printing a separate code.

The full packaging-to-browser chain, including what happens at the factory and in the scanner, is followed step by step in how a GS1 Digital Link works, from packaging to browser.

Link types: one product, many destinations

A resolver can hold many links for the same item. Each link has three parts: a target (the web address it points to), a link type (what kind of resource it is) and a title (a human-readable label).

Link types come from the GS1 Web Vocabulary and are written in a short form such as gs1:pip. A client — an app, a partner system, or a specially formed link on a web page — can ask for a particular type by adding it to the request, for example ?linkType=gs1:instructions.

Some of the link types GS1 defines:

Because the types are shared vocabulary, a retailer's app or a recycling service can ask any conformant resolver for the same kind of resource and get a predictable answer. GS1 in Europe has also proposed GS1 Digital Link as one way to connect products to an EU Digital Product Passport, although EU rules do not require it. How brands use this to serve different audiences from one code is covered in how one product identity can lead to product info, warranty, loyalty and traceability.

Falling back up the tree

Brands rarely want to register links for every single serial number. The standard handles this with inheritance: links are inherited "up the tree". If a request names a serial number and nothing has been registered for that specific serial, the resolver falls back to the links registered at a less specific level, such as the GTIN.

In practice, Maison Ardent registers its product page and instructions once, against the GTIN. Every serialized bottle of that perfume inherits them. If one batch is later affected by a quality issue, the brand can register a different link against that batch only, and the rest of the range is unaffected.

When the resolver cannot pick one link

Two situations are defined clearly:

• No match. If a request asks for a link type the item does not have, the resolver returns a 404 ("not found") response.
• Several equal matches. If more than one link fits the request equally well, the resolver returns a 300 ("multiple choices") response, letting the client choose.

These sound like technical details, but they are what make a resolver predictable for machines. A partner's system can ask for recall status and know the difference between "there is no recall link" and "there are two options, pick one".

The linkset: the full directory

Sometimes a client does not want to be redirected at all. It wants to know everything available for an item. For that, a resolver can return a linkset: the complete list of links for the item, in a standard format defined by RFC 9264.

A client asks for it either by requesting linkType=linkset or by sending the header Accept: application/linkset+json. When a linkset is requested, the resolver does not redirect; it returns the list as data. That makes a resolver discoverable: software can find out what a product offers without guessing addresses.

Who runs a resolver?

Whoever controls the domain printed in the code controls what happens when it is scanned. GS1 recommends that brands use their own domain. A brand can run the resolver itself or use a service provider, but the domain decision is long-lived because the address is printed on packs. We look at that choice in why brands should separate product identity from the digital experience.

GS1 itself operates id.gs1.org, which describes itself as "The GS1 AISBL Resolver". Its published description lists the primary keys it supports, including GTIN (01), SSCC (00) and GLN (414), and the canonical form of a GS1 Digital Link uses that domain. The resolver standard names it as the GS1 Global Office's resolver but does not describe it as a host for brands' consumer content, so do not plan on it serving your product pages.

GS1's retail 2D guideline describes a network of conformant resolvers acting as a discovery service, where product information "remains within [the] data owner's own system". In other words, a resolver tells you where information lives; it is not a central database of it.

What a resolver does not do

Three other boundaries are worth stating plainly:

• It does not prove authenticity. A resolver can answer for any well-formed address, including one copied from a genuine pack. Even the gs1:verificationService link type points to a service that checks the status of an identifier, not whether a physical product is genuine. Read why a QR code does not automatically prove authenticity.
• It does not store supply-chain history. Event records belong in EPCIS systems run by the businesses that capture them. A resolver can point to a brand's EPCIS source through gs1:epcis, but it does not hold the events. See what EPCIS is.
• It is not the content. The resolver holds links. The product page, the warranty form and the loyalty app are separate systems, which is precisely why they can change independently.

How the pieces fit together

A useful way to remember the resolver's role is as a receptionist for the product: it knows which desk handles which request and sends each visitor to the right one. We develop that analogy, and where it breaks down, in think of a resolver as a receptionist for a product.

Is a simple redirect enough?

Not every brand needs every capability above on day one. A single product page per GTIN, reached by a plain web redirect, can be a sensible start. The difference shows up when you need several destinations, partner access, fallbacks or machine-readable discovery. We compare the two directly in resolver vs redirect.

The short version

A GS1 resolver is the part of a connected product that can change after printing. The code fixes what the item is; the resolver decides where that identity leads, with one default link for ordinary scans, typed links for specific needs, sensible fallbacks, and a full linkset for software that wants the whole picture. Get the identity and the domain right at print time, and the resolver gives you room to change everything else later.