JSON has proven to be a highly useful object serialization and messaging format. In an attempt to harmonize the representation of Linked Data in JSON, this specification outlines a common JSON representation format for expressing directed graphs; mixing both Linked Data and non-Linked Data in a single document.

This document has been under development for over 25 months in the JSON for Linking Data Community Group. The document has recently been transferred to the RDF Working Group for review, improvement, and publication. The specification has undergone significant development, review, and changes during the course of the last 25 months.

There are several independent interoperable implementations of this specification. There is a fairly complete test suite and a live JSON-LD editor that is capable of demonstrating the features described in this document. While development on implementations, the test suite and the live editor will continue, they are believed to be mature enough to be integrated into a non-production system at this point in time with the expectation that they could be used in a production system within the next year.

There are a number of ways that one may participate in the development of this specification:


Linked Data is a technique for creating a network of inter-connected data across different documents and Web sites. In general, Linked Data has four properties: 1) it uses IRIs to name things; 2) it uses HTTP IRIs for those names; 3) the name IRIs, when dereferenced, provide more information about the name; and 4) the data expresses links to data on other Web sites. These properties allow data published on the Web to work much like Web pages do today. One can start at one piece of Linked Data, and follow the links to other pieces of data that are hosted on different sites across the Web.

JSON-LD is designed as a lightweight syntax to express Linked Data in JSON [[!RFC4627]]. It is primarily intended to be a way to use Linked Data in Web-based programming environments. It is also useful when building interoperable Web services and when storing Linked Data in JSON-based storage engines. Since JSON-LD is 100% compatible with JSON the large number of JSON parsers and libraries available today can be reused. Additionally to all the features JSON provides, JSON-LD introduces:

Developers that require any of the facilities listed above or need to serialize an RDF graph or dataset [[RDF-CONCEPTS]] in a JSON-based syntax will find JSON-LD of interest. The syntax is designed to not disturb already deployed systems running on JSON, but provide a smooth upgrade path from JSON to JSON-LD.

How to Read this Document

This document is a detailed specification for a serialization of Linked Data in JSON. The document is primarily intended for the following audiences:

This specification does not describe the programming interfaces for the JSON-LD Syntax. The specification that describes the programming interfaces for JSON-LD documents is the JSON-LD Application Programming Interface [[JSON-LD-API]].

To understand the basics in this specification you must first be familiar with JSON, which is detailed in [[!RFC4627]].

Design Goals and Rationale

A number of design goals were established before the creation of this markup language:

No extra processors or software libraries should be necessary to use JSON-LD in its most basic form. The language will provide developers with a very easy learning curve. Developers only need to know JSON and two keywords (@context and @id) to use the basic functionality in JSON-LD.
The JSON-LD markup must be 100% compatible with JSON. This ensures that all of the standard JSON libraries work seamlessly with JSON-LD documents.
The syntax must be able to express directed graphs, which have been proven to be able to express almost every real world data model.
The JSON-LD syntax must be very terse and human readable, requiring as little effort as possible from the developer.
Zero Edits, most of the time
JSON-LD must make the transition to JSON-LD as simple as possible. In many cases, zero edits to the JSON document and the addition of one line to the HTTP response should suffice (see ). This allows organizations that have already deployed large JSON-based infrastructure to use JSON-LD's features in a way that is not disruptive to their day-to-day operations and is transparent to their current customers. However, there are times where mapping JSON to a graph representation is more complex than a simple one-line change. In these instances, rather than extending JSON-LD to support an esoteric use case, we chose not to support the use case. While Zero Edits is a design goal, it is not always possible without adding great complexity to the language. We should focus on simplicity when possible.


General Terminology

This document uses the following terms as defined in JSON [[!RFC4627]]. Refer to the JSON Grammar section in [[!RFC4627]] for formal definitions.

JSON object
An object structure is represented as a pair of curly brackets surrounding zero or more key-value pairs. A key is a string. A single colon comes after each key, separating the key from the value. A single comma separates a value from a following key.
An array structure is represented as square brackets surrounding zero or more values. Values are separated by commas. In JSON, an array is an ordered sequence of zero or more values. While JSON-LD uses the same array representation as JSON, the collection is unordered by default. While order is preserved in regular JSON arrays, it is not in regular JSON-LD arrays unless specific markup is provided (see ).
A string is a sequence of zero or more Unicode characters, wrapped in double quotes, using backslash escapes (if necessary). A character is represented as a single character string.
A number is similar to that used in most programming languages, except that the octal and hexadecimal formats are not used and that leading zeros are not allowed.
true and false
Values that are used to express one of two possible boolean states.
The null value, which is typically used to clear or forget data. For example, A key-value pair in the @context where the value is null explicitly decouples a term's association with an IRI. A key-value pair in the body of a JSON-LD document whose value is null has the same meaning as if the key-value pair was not defined. If @value, @list, or @set is set to null in expanded form, then the entire JSON object is ignored.

Syntax Tokens and Keywords

JSON-LD specifies a number of syntax tokens and keywords that are a core part of the language:

Used to define the short-hand names that are used throughout a JSON-LD document. These short-hand names are called terms and help developers to express specific identifiers in a compact manner. The @context keyword is described in detail in the section titled .
Used to uniquely identify things that are being described in the document. This keyword is described in .
Used to specify the data that is associated with a particular property in the graph. This keyword is described in and .
Used to specify the native language for a particular value or the default language of a JSON-LD document. This keyword is described in the section titled .
Used to set the data type of a node or typed value. This keyword is described in the section titled .
Used to set the default container type for a term. This keyword is described in the section titled .
Used to express an ordered set of data. This keyword is described in the section titled .
Used to express an unordered set of data and to ensure that values are always represented as arrays. This keyword is described in the section titled .
Used to specify that a container is used to index information and that processing should continue deeper into a JSON data structure. This keyword is described in the section titled .
Used to expand properties and values in @type with a common prefix IRI. This keyword is described in section .
Used to explicitly label a JSON-LD graph. This keyword is described in .
The separator for JSON keys and values that use compact IRIs.

For the avoidance of doubt, all keys, keywords, and values in JSON-LD are case-sensitive.


This specification describes the conformance criteria for JSON-LD documents. This criteria is relevant to authors and authoring tool implementers.

A JSON-LD document complies with this specification if it follows the normative statements in section . JSON documents can be interpreted as JSON-LD by following the normative statements in section . For convenience, normative statements for documents are often phrased as statements on the properties of the document.

The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, NOT RECOMMENDED, MAY, and OPTIONAL in this specification have the meaning defined in [[!RFC2119]].

Basic Concepts

JSON [[RFC4627]] is a lightweight, language-independent data-interchange format. It is easy to parse and easy to generate. However, it is difficult to integrate JSON from different sources as the data has just local meaning. Furthermore, JSON has no built-in support for hyperlinks - a fundamental building block on the Web. Let's look at an example that we will be using for the rest of this section:


It's obvious for humans that the data is about a person whose name is "Manu Sporny" and that the homepage property contains the URL of that person's homepage. A machine doesn't have such an intuitive understanding and sometimes, even for humans, it is difficult to resolve ambiguities in such representations. This problem can be solved by using unambiguous identifiers to denote the different concepts instead of terms such as "name", "homepage", etc.

Linked Data, and the Web in general, uses IRIs (Internationalized Resource Identifiers as described in [[!RFC3987]]) for unambiguous identification. The idea is to assign IRIs to something that may be of use to other developers and that it is useful to give them an unambiguous identifier. That is, it is useful for terms to expand to IRIs so that developers don't accidentally step on each other's terms. Furthermore, developers and machines are able to use this IRI (by using a web browser, for instance) to go to the term and get a definition of what the term means.

Leveraging the well-known schema.org vocabulary, the example above could be unambiguously expressed as follows:


In the example above, every property is unambiguously identified by an IRI and all values representing IRIs are explicitly marked as such by the @id keyword. While this is a valid JSON-LD document that is very specific about its data, the document is also overly verbose and difficult to work with for human developers. To address this issue, JSON-LD introduces the notion of a context as described in the next section.

The Context

Simply speaking, a context is used to map terms, i.e., properties with associated values, to IRIs. Terms are case sensitive and any valid string that is not a reserved JSON-LD keyword can be used as a term.

For the sample document in the previous section, a context would look something like this:


As the context above shows, the value of a term definition can either be a simple string, mapping the term to an IRI, or a JSON object.

When a JSON object is associated with a term, it is called an expanded term definition. Expanded term definitions may be used to associate type or language information with a term. The example above specifies that the values of image and homepage terms are IRIs. They also allow terms to be used for index maps and to specify whether array values are to be interpreted as sets or lists. Expanded term definitions may be defined using absolute or compact IRIs as keys, which is mainly used to associate type or language information with an absolute or compact IRI.

Contexts can either be directly embedded into the document or be referenced. Assuming the context document in the previous example can be retrieved at http://json-ld.org/contexts/person.jsonld, it can be referenced by adding a single line and allows a JSON-LD document to be expressed much more concisely as shown in the example below:


The referenced context not only specifies how the terms map to IRIs in the Schema.org vocabulary but also specifies that the values of the homepage and image property can be interpreted as an IRI ("@type": "@id", see section for more details). This information gives the data global context and allows developers to re-use each other's data without having to agree to how their data will interoperate on a site-by-site basis. External JSON-LD context documents may contain extra information located outside of the @context key, such as documentation about the terms declared in the document. Information contained outside of the @context value is ignored when the document is used as an external JSON-LD context document.

Contexts may also be specified in-line. This has the advantage that JSON-LD documents can be processed even in the absence of a connection to the Web.



IRIs (Internationalized Resource Identifiers [[!RFC3987]]) are fundamental to Linked Data as that is how most nodes and properties are identified. In JSON-LD, IRIs may be represented as an absolute IRI or a relative IRI. An absolute IRI is defined in [[!RFC3987]] as containing a scheme along with path and optional query and fragment segments. A relative IRI is an IRI that is relative to some other absolute IRI. In JSON-LD all relative IRIs are resolved relative to the base IRI associated with the document, which is typically the directory path containing the document.

IRIs can be expressed directly in the key position like so:


In the example above, the key http://schema.org/name is interpreted as an absolute IRI because it contains a colon (:) and the "http" prefix does not exist in the context.

Term-to-IRI expansion occurs if the key matches a term defined within the active context:


JSON keys that do not expand to an absolute IRI are ignored, or removed in some cases, by the [[JSON-LD-API]]. However, JSON keys that do not include a mapping in the context are still considered valid expressions in JSON-LD documents—the keys just don't expand to unambiguous identifiers.

At times, all properties and types may come from the same vocabulary. JSON-LD's @vocab keyword allows an author to set a common prefix to be used for all properties and types that do not match a term or are neither a compact IRI nor an absolute IRI (i.e., they do not contain a colon).


An IRI is generated when a JSON object is used in the value position and contains an @id keyword:


Specifying a JSON object with an @id key is used to identify that node using an IRI. This facility may also be used to link to another node object using a mechanism called embedding, which is covered in the section titled .

If type coercion rules are specified in the @context for a particular term or property IRI, an IRI is generated:


In the example above, even though the value http://manu.sporny.org/ is expressed as a JSON string, the type coercion rules will transform the value into an IRI when generating the JSON-LD graph. See for more details about this feature.

In summary, IRIs can be expressed in a variety of different ways in JSON-LD:

  1. JSON object keys that have a term mapping in the active context expand to an IRI (only applies outside of the context definition).
  2. If there is a @vocab mapping in the active context, JSON object keys without an explicit mapping in the active context are expanded to an IRI.
  3. An IRI is generated for the string value specified using @id or @type.
  4. An IRI is generated for the string value of any key for which there are coercion rules that contain a @type key that is set to a value of @id or @vocab.

Node Identifiers

To be able to externally reference nodes in a graph, it is important that each node has an unambiguous identifier. IRIs are a fundamental concept of Linked Data, and nodes should have a de-referenceable identifier used to name and locate them. For nodes to be truly linked, de-referencing the identifier should result in a representation of that node. Associating an IRI with a node tells an application that it can fetch the resource associated with the IRI and get back a description of the node.

JSON-LD documents may also contain descriptions of other nodes, so it is necessary to be able to uniquely identify each node so that the data is associated with the correct node in an unambiguous way.

A node is identified using the @id keyword:


The example above contains a node object identified by the IRI http://example.org/people#joebob.

Specifying the Type

The type of a particular node can be specified using the @type keyword. In Linked Data, types are uniquely identified with an IRI.

A node can be assigned more than one type by using an array:

The value of a @type key may also be a term defined in the active context:

Advanced Concepts

This section is normative.

JSON-LD has a number of features that provide functionality above and beyond the core functionality described above. The following section describes this advanced functionality in more detail.

Compact IRIs

A document on the Web that defines one or more IRIs for use as properties in Linked Data is called a vocabulary. Terms in Linked Data documents may draw from a number of different vocabulariess. At times, declaring every single term that a document uses can require the developer to declare tens, if not hundreds of potential vocabulary terms that are used across an application. This is a concern for at least two reasons: the first is the cognitive load on the developer of remembering all of the terms, and the second is the serialized size of the context if it is specified inline. In order to address these issues, the concept of a compact IRI is introduced.

A compact IRI is a way of expressing an IRI using a prefix and suffix separated by a colon (:) which is similar to the CURIE Syntax in [[RDFA-CORE]]. The prefix is a term taken from the active context and is a short string identifying a particular IRI in a JSON-LD document. For example, the prefix foaf may be used as a short hand for the Friend-of-a-Friend vocabulary, which is identified using the IRI http://xmlns.com/foaf/0.1/. A developer may append any of the FOAF vocabulary terms to the end of the prefix to specify a short-hand version of the absolute IRI for the vocabulary term. For example, foaf:name would be expanded out to the IRI http://xmlns.com/foaf/0.1/name. Instead of having to remember and type out the entire IRI, the developer can instead use the prefix in their JSON-LD markup.

Prefixes are expanded when the form of the value is a compact IRI represented as a prefix:suffix combination, and the prefix matches a term defined within the active context:


foaf:name above will automatically expand out to the IRI http://xmlns.com/foaf/0.1/name.

Terms are interpreted as compact IRIs if they contain at least one colon and the first colon is not followed by two slashes (//, as in http://example.com). To generate the full IRI, the value is first split into a prefix and suffix at the first occurrence of a colon (:). If the active context contains a term mapping for prefix, an IRI is generated by prepending the mapped prefix to the (possibly empty) suffix using textual concatenation. If no prefix mapping is defined, the value is interpreted as an absolute IRI. If the prefix is an underscore (_), the IRI remains unchanged.

Consider the following example:


In this example, two different vocabularies are referred to using prefixes. Those prefixes are then used as type and property values using the compact IRI prefix:suffix notation.

It's also possible to use compact IRIs within the context as shown in the following example:


Typed Values

A value with an associated type, also known as a typed value, is indicated by associating a value with an IRI which indicates the value's type. Typed values may be expressed in JSON-LD in three ways:

  1. By utilizing the @type keyword when defining a term within a @context section.
  2. By utilizing an expanded typed value.
  3. By using a native JSON type such as number, true, or false.

The first example uses the @type keyword to associate a type with a particular term in the @context:

The modified key's value above is automatically type coerced to a dateTime value because of the information specified in the @context. A JSON-LD processor will interpret the markup above like so:

Subject Property Value Value Type
http://example.com/docs/1 http://purl.org/dc/terms/modified 2010-05-29T14:17:39+02:00 http://www.w3.org/2001/XMLSchema#dateTime

The second example uses the expanded form of setting the type information in the body of a JSON-LD document:

Both examples above would generate the value 2010-05-29T14:17:39+02:00 with the type http://www.w3.org/2001/XMLSchema#dateTime. Note that it is also possible to use a term or a compact IRI to express the value of a type.

The @type keyword is also used to associate a type with a node. The concept of a node type and a value type are different.

Generally speaking, a node type specifies the type of thing that is being described, like a person, place, event, or web page. A value type specifies the unit of measurement for a particular value, such as a date, meter, or light year.

The first use of @type associates a node type (http://schema.org/BlogPosting) with the node, which is expressed using the @id keyword. The second use of @type associates a value type (http://www.w3.org/2001/XMLSchema#dateTime) with the value expressed using the @value keyword. As a general rule, when @value and @type are used in the same JSON object, the @type keyword is expressing a value type. Otherwise, the @type keyword is expressing a node type. The markup above expresses the following data:

Subject Property Value Value Type
http://example.org/posts#TripToWestVirginia http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://schema.org/BlogPosting -
http://example.org/posts#TripToWestVirginia http://purl.org/dc/terms/modified 2010-05-29T14:17:39+02:00 http://www.w3.org/2001/XMLSchema#dateTime

Type Coercion

JSON-LD supports the coercion of values to particular data types. Type coercion allows someone deploying JSON-LD to coerce the incoming or outgoing values to the proper data type based on a mapping of data type IRIs to terms. Using type coercion, value representation is preserved without requiring the data type to be specified with each piece of data.

Type coercion is specified within an expanded term definition using the @type key. The value of this key expands to an IRI. Alternatively, the keywords @id or @vocab may be used as value to indicate that within the body of a JSON-LD document, a string value of a term coerced to @id or @vocab is to be interpreted as an IRI. The difference between @id and @vocab is how values are expanded to absolute IRIs. @vocab first tries to expand the value by interpreting it as term. If no matching term is found in the active context, it tries to expand it as compact IRI or absolute IRI if there's a colon in the value; otherwise, it will expand the value using the active context's vocabulary mapping, if present, or by interpreting it as relative IRI. Values coerced to @id in contrast are expanded as compact IRI or absolute IRI if a colon is present; otherwise, they are interpreted as relative IRI.

Terms or compact IRIs used as the value of a @type key may be defined within the same context. This means that one may specify a term like xsd and then use xsd:integer within the same context definition.

The example below demonstrates how a JSON-LD author can coerce values to typed values, IRIs, and lists.

The markup shown above would generate the following data. The data has no inherent order except for the values of the http://schema.org/homepage property which represent an ordered list.

Subject Property Value Value Type
http://example.com/people#john http://xmlns.com/foaf/0.1/name John Smith  
http://example.com/people#john http://xmlns.com/foaf/0.1/age 41 http://www.w3.org/2001/XMLSchema#integer
http://example.com/people#john http://xmlns.com/foaf/0.1/homepage http://personal.example.org/  

Terms may also be defined using absolute IRIs or compact IRIs. This allows coercion rules to be applied to keys which are not represented as a simple term. For example:

In this case the @id definition in the term definition is optional, but if it does exist, the compact IRI or IRI is treated as a term (not a prefix:suffix construct) so that the actual definition of a prefix becomes unnecessary. Type coercion is performed using the unexpanded value of the key if there is an exact match for the key in the active context.

Keys in the context are treated as terms for the purpose of expansion and value coercion. At times, this may result in multiple representations for the same expanded IRI. For example, one could specify that dog and cat both expanded to http://example.com/vocab#animal. Doing this could be useful for establishing different type coercion or language specification rules. It also allows a compact IRI (or even an absolute IRI) to be defined as something else entirely. For example, one could specify that the term http://example.org/zoo should expand to http://example.org/river, but this usage is discouraged because it would lead to a great deal of confusion among developers attempting to understand the JSON-LD document.

Advanced Context Usage

Section introduced the basics of what makes JSON-LD work. This section expands on the basic principles of the context and demonstrates how more advanced use cases can be achieved using JSON-LD.

In general, contexts may be used at any time a JSON object is defined. The only time that one cannot express a context is inside a context definition itself. For example, a JSON-LD document may use more than one context at different points in a document:


Duplicate context terms are overridden using a last-defined-wins mechanism.


In the example above, the name term is overridden in the more deeply nested details structure. Note that this is rarely a good authoring practice and is typically used when working with legacy applications that depend on a specific structure of the JSON object. If a term is redefined within a context, all previous rules associated with the previous definition are removed. If a term is redefined to null, the term is effectively removed from the list of terms defined in the active context.

Multiple contexts may be combined using an array, which is processed in order. The set of contexts defined within a specific JSON object are referred to as local contexts. The active context refers to the accumulation of local contexts that are in scope at a specific point within the document. Setting a local context to null effectively resets the active context to an empty context. The following example specifies an external context and then layers an embedded context on top of the external context:


It is a best practice to put the context definition at the top of the JSON-LD document.

To avoid forward-compatibility issues, terms starting with an @ character are to be avoided as they might be used as keywords in future versions of JSON-LD. Furthermore, the use of empty terms ("") is discouraged as not all programming languages are able to handle empty property names.

Interpreting JSON as JSON-LD

Ordinary JSON documents can be interpreted as JSON-LD by referencing a JSON-LD context document in an HTTP Link Header. Doing so allows JSON to be unambiguously machine-readable without requiring developers to drastically change their markup and provides an upgrade path for existing infrastructure without breaking existing clients that rely on the application/json media type.

In order to use an external context with an ordinary JSON document, an author MUST specify an IRI to a valid JSON-LD document in an HTTP Link Header [[!RFC5988]] using the http://www.w3.org/ns/json-ld#context link relation. The referenced document MUST have a top-level JSON object. The @context subtree within that object is added to the top-level JSON object of the referencing document. If an array is at the top-level of the referencing document and its items are JSON objects, the @context subtree is added to all array items. All extra information located outside of the @context subtree in the referenced document MUST be discarded. Effectively this means that the active context is initialized with the referenced external context.

The following example demonstrates the use of an external context with an ordinary JSON document:


Please note that JSON-LD documents served with the application/ld+json media type MUST have all context information, including references to external contexts, within the body of the document. Contexts linked via a http://www.w3.org/ns/json-ld#context HTTP Link Header MUST be ignored for such documents.

String Internationalization

At times, it is important to annotate a string with its language. In JSON-LD this is possible in a variety of ways. First, it is possible to define a default language for a JSON-LD document by setting the @language key in the context:


The example above would associate the ja language code with the two strings 花澄 and 科学者. Languages codes are defined in [[!BCP47]].

To clear the default language for a subtree, @language can be set to null in a local context as follows:


Second, it is possible to associate a language with a specific term using an expanded term definition:


The example above would associate 忍者 with the specified default language code ja, Ninja with the language code en, and Nindža with the language code cs. The value of name, Yagyū Muneyoshi wouldn't be associated with any language code since @language was reset to null in the expanded term definition.

Language associations can only be applied to plain literal strings. Typed values or values that are subject to cannot be language tagged.

Just as in the example above, systems often need to express the value of a property in multiple languages. Typically, such systems also try to ensure that developers have a programmatically easy way to navigate the data structures for the language-specific data. In this case, language maps may be utilized.


The example above expresses exactly the same information as the previous example but consolidates all values in a single property. To access the value in a specific language in a programming language supporting dot-notation accessors for object properties, a developer may use the property.language pattern. For example, to access the occupation in English, a developer would use the following code snippet: obj.occupation.en.

Third, it is possible to override the default language by using an expanded value:


This makes it possible to specify a plain string by omitting the @language tag or setting it to null when expressing it using an expanded value:


Overriding @vocab

If @vocab is used but certain keys in an object should not be expanded using the vocabulary IRI, a term can be explicitly set to null in the context. For instance, in the example below the databaseId member would be ignored by a JSON-LD processor.


Property Generators

At times, an author may find that they need to express the same value for multiple properties. The simplest approach to accomplish this goal would be to do the following:

Unfortunately, the approach above produces redundant data and would become a publishing burden for large data sets. In these situations, the author may use a property generator to express a term that maps to multiple properties in the JSON-LD graph. This method can be accomplished by using the following markup pattern:

While the term above is only used once outside of the @context, the document above will be interpreted like so:

Subject Property Value
http://example.com/book http://purl.org/dc/terms/title The Count of Monte Cristo
http://example.com/book http://schema.org/name The Count of Monte Cristo
http://example.com/book http://www.w3.org/2000/01/rdf-schema#label The Count of Monte Cristo

IRI Expansion Within a Context

In general, normal IRI expansion rules apply anywhere an IRI is expected (see ). Within a context definition, this can mean that terms defined within the context may also be used within that context as long as there are no circular dependencies. For example, it is common to use the xsd namespace when defining typed values:

In this example, the xsd term is defined and used as a prefix for the @type coercion of the age property.

Terms may also be used when defining the IRI of another term:

Compact IRIs and IRIs may be used on the left-hand side of a term definition.

In this example, the compact IRI form is used in two different ways. In the first approach, foaf:age declares both the IRI for the term (using short-form) as well as the @type associated with the term. In the second approach, only the @type associated with the term is specified. The full IRI for foaf:homepage is determined by looking up the foaf prefix in the context.

Absolute IRIs may also be used in the key position in a context:

In order for the absolute IRI to match above, the absolute IRI needs to be used in the JSON-LD document. Also note that foaf:homepage will not use the { "@type": "@id" } declaration because foaf:homepage is not the same as http://xmlns.com/foaf/0.1/homepage. That is, terms are looked up in a context using direct string comparison before the prefix lookup mechanism is applied.

While it is possible to define a compact IRI, or an absolute IRI to expand to some other unrelated IRI (for example, foaf:name expanding to http://example.org/unrelated#species), such usage is strongly discouraged.

The only exception for using terms in the context is that circular definitions are not allowed. That is, a definition of term1 cannot depend on the definition of term2 if term2 also depends on term1. For example, the following context definition is illegal:

Sets and Lists

A JSON-LD author can express multiple values in a compact way by using arrays. Since graphs do not describe ordering for links between nodes, arrays in JSON-LD do not provide an ordering of the contained elements by default. This is exactly the opposite from regular JSON arrays, which are ordered by default. For example, consider the following simple document:

The markup shown above would result in the following data being generated, each relating the node to an individual value, with no inherent order:

Subject Property Value
http://example.org/people#joebob http://xmlns.com/foaf/0.1/nick joe
http://example.org/people#joebob http://xmlns.com/foaf/0.1/nick bob
http://example.org/people#joebob http://xmlns.com/foaf/0.1/nick JB

Multiple values may also be expressed using the expanded form:

The markup shown above would generate the following data, again with no inherent order:

Subject Property Value Language
http://example.org/articles/8 http://purl.org/dc/terms/title Das Kapital de
http://example.org/articles/8 http://purl.org/dc/terms/title Capital en

As the notion of ordered collections is rather important in data modeling, it is useful to have specific language support. In JSON-LD, a list may be represented using the @list keyword as follows:

This describes the use of this array as being ordered, and order is maintained when processing a document. If every use of a given multi-valued property is a list, this may be abbreviated by setting @container to @list in the context:

List of lists are not allowed in this version of JSON-LD. This decision was made due to the extreme amount of added complexity when processing lists of lists.

While @list is used to describe ordered lists, the @set keyword is used to describe unordered sets. The use of @set in the body of a JSON-LD document is optimized away when processing the document, as it is just syntactic sugar. However, @set is helpful when used within the context of a document. Values of terms associated with a @set or @list container are always represented in the form of an array, even if there is just a single value that would otherwise be optimized to a non-array form in compact form (see ). This makes post-processing of JSON-LD documents easier as the data is always in array form, even if the array only contains a single value.

The use of @container in the body of a JSON-LD document has no meaning and is not allowed by the JSON-LD grammar (see ).


Embedding is a JSON-LD feature that allows an author to use node objects as property values. This is a commonly used mechanism for creating a parent-child relationship between two nodes.

The example shows two nodes related by a property from the first node:


A node object, like the one used above, may be used in any value position in the body of a JSON-LD document.

Named Graphs

At times, it is necessary to make statements about a JSON-LD graph itself, rather than just a single node. This can be done by grouping a set of nodes using the @graph keyword. A developer may also name data expressed using the @graph keyword by pairing it with an @id keyword as shown in the following example:


The example above expresses a named JSON-LD graph that is identified by the IRI http://example.org/graphs/73. That graph is composed of the statements about Manu and Gregg. Metadata about the graph itself is also expressed via the generatedAt property, which specifies when the graph was generated. An alternative view of the information above is represented in table form below:

Graph Subject Property Value Value Type
http://example.org/graphs/73 http://example.org/graphs/73 http://www.w3.org/ns/prov#generatedAtTime 2012-04-09 http://www.w3.org/2001/XMLSchema#date
http://example.org/graphs/73 http://manu.sporny.org/i/public http://www.w3.org/2001/XMLSchema#type http://xmlns.com/foaf/0.1/Person
http://example.org/graphs/73 http://manu.sporny.org/i/public http://xmlns.com/foaf/0.1/name Manu Sporny
http://example.org/graphs/73 http://manu.sporny.org/i/public http://xmlns.com/foaf/0.1/knows http://greggkellogg.net/foaf#me
http://example.org/graphs/73 http://greggkellogg.net/foaf#me http://www.w3.org/2001/XMLSchema#type http://xmlns.com/foaf/0.1/Person
http://example.org/graphs/73 http://greggkellogg.net/foaf#me http://xmlns.com/foaf/0.1/name Gregg Kellogg
http://example.org/graphs/73 http://greggkellogg.net/foaf#me http://xmlns.com/foaf/0.1/knows http://manu.sporny.org/i/public

When a JSON-LD document's top-level structure is an object that contains no other properties than @graph and optionally @context (properties that are not mapped to an IRI or a keyword are ignored), @graph is considered to express the otherwise implicit default graph. This mechanism can be useful when a number of nodes exist at the document's top level that share the same context. The @graph keyword collects such nodes in an array and allows the use of a shared context.


In this case, embedding doesn't work as each node object references the other. This is equivalent to using multiple node objects in array and defining the @context within each node object:


Identifying Blank Nodes

At times, it becomes necessary to be able to express information without being able to uniquely identify the node. This type of node is called a blank node (see Section 3.4: Blank Nodes of [[RDF-CONCEPTS]]). In JSON-LD, blank node identifiers are automatically created if an IRI is not specified using the @id keyword. However, authors may provide identifiers for blank nodes by using the special _ (underscore) prefix. This allows one to reference the node locally within the document, but makes it impossible to reference the node from an external document. The blank node identifier is scoped to the document in which it is used.


The example above would set the node to _:foo, which can then be used elsewhere in the JSON-LD document to refer back to the blank node. If a developer finds that they refer to the blank node more than once, they should consider naming the node using a dereferenceable IRI so that it can also be referenced from other documents.

Aliasing Keywords

Each of the JSON-LD keywords, except for @context, may be aliased to application-specific keywords. This feature allows legacy JSON content to be utilized by JSON-LD by re-using JSON keys that already exist in legacy documents. This feature also allows developers to design domain-specific implementations using only the JSON-LD context.


In the example above, the @id and @type keywords have been given the aliases url and a, respectively.

Since keywords cannot be redefined, they can also not be aliased to other keywords. Every statement in the context having a keyword as the key (as in { "@type": ... }) will be ignored when being processed.

Expanded Document Form

The JSON-LD Algorithms and API specification [[JSON-LD-API]] defines a method for expanding a JSON-LD document. Expansion is the process of taking a JSON-LD document and applying a @context such that all IRIs, types, and values are expanded so that the @context is no longer necessary.

For example, assume the following JSON-LD input document:

Running the JSON-LD Expansion algorithm against the JSON-LD input document provided above would result in the following output:

Expanded document form is useful when an application has to process input data in a deterministic form. It has been optimized to ensure that the code that developers have to write is minimized compared to the code that would have to be written to operate on .

Compact Document Form

The JSON-LD Algorithms and API specification [[JSON-LD-API]] defines a method for compacting a JSON-LD document. Compaction is the process of taking a JSON-LD document and applying a context such that a very compact form of the document is generated. At times, a JSON-LD document may be received that is not in its most compact form. The JSON-LD Algorithms, via an API, provides a way to compact a JSON-LD document.

For example, assume the following JSON-LD input document:

Additionally, assume the following developer-supplied JSON-LD context:

Running the JSON-LD Compaction algorithm given the context supplied above against the JSON-LD input document provided above would result in the following output:

The compaction algorithm enables a developer to map any document into an application-specific compacted form. The process consists of expanding the document (see ) and then using a developer-supplied context to compact the expanded document. While the context provided above mapped http://xmlns.com/foaf/0.1/name to name, it could have also mapped it to any arbitrary term provided by the developer. This powerful mechanism allows the developer to re-shape the incoming JSON data into a format that is optimized for their application.

Data Indexing

Databases are typically used to make access to data more efficient. Developers often extend this sort of functionality into their application data to deliver similar performance gains. Often this data does not have any meaning from a Linked Data standpoint, but is still useful for an application.

JSON-LD introduces the notion of index maps that can be used to structure data into a form that is more efficient to access. The data indexing feature allows an author to structure data using a simpley key-value map where the keys do not map to IRIs. This enables direct access to data instead of having to scan an array in search of a specific item. In JSON-LD such data can be specified by associating the @index keyword with a @container declaration in the context:


In the example above, the blogPost term has been marked as an index map. The en, de, and ja keys will be ignored semantically, but preserved syntactically, by the JSON-LD Processor. This allows a developer to access the German version of the blogPost using the following code snippet: obj.blogPost.de.

The interpretation of the data above is expressed in the table below. Note how the index keys do not appear in the Linked Data below, but would continue to exist if the document were compacted or expanded (see and ) using a JSON-LD processor:

Subject Property Value
http://example.com/ http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://schema.org/Blog
http://example.com/ http://schema.org/name World Financial News
http://example.com/ http://schema.org/blogPost http://example.com/posts/1/en
http://example.com/ http://schema.org/blogPost http://example.com/posts/1/de
http://example.com/posts/1/en http://schema.org/articleBody World commodities were up today with heavy trading of crude oil...
http://example.com/posts/1/en http://schema.org/wordCount 1539
http://example.com/posts/1/de http://schema.org/articleBody Die Werte an Warenbörsen stiegen im Sog eines starken Handels von Rohöl...
http://example.com/posts/1/de http://schema.org/wordCount 1204

Data Model

JSON-LD is a serialization format for Linked Data based on JSON. It is therefore important to distinguish between the syntax, which is defined by JSON in [[RFC4627]], and JSON-LD's data model which is defined as follows:

In contrast to the RDF data model as defined in [[RDF-CONCEPTS]], JSON-LD allows blank nodes as property labels and graph names. This feature is controversial in the RDF WG and may be removed in the future.

JSON-LD documents MAY contain data that cannot be represented by the data model defined above. Unless otherwise specified, such data is ignored when a JSON-LD document is being processed. This means, e.g., that properties which are not mapped to an IRI or blank node will be ignored.

Figure 1: An illustration of JSON-LD's data model.

JSON-LD Grammar

This section is normative

This appendix restates the syntactic conventions described in the previous sections more formally.

The JSON-LD context allows keywords ). Whenever a keyword is discussed in this grammar, the statements also apply to an alias for that keyword. For example, if the active context defines the term id as an alias for @id, that alias may be legitimately used as a substitution for @id. Note that keyword aliases are not expanded during context processing.

A JSON-LD document MUST be a valid JSON document as described in [[!RFC4627]].

A JSON-LD document MUST be a single node object or a JSON array containing a set of one or more node objects.

Node Object

A node object represents zero or more properties of a node in the JSON-LD graph serialized by the JSON-LD document. A JSON object is a node object if it exists outside of a JSON-LD context and:

The properties of a node in a JSON-LD graph may be spread among different node objects within a document. When that happens, the keys of the different node objects are merged to create the properties of the resulting node.

A node object MUST be a JSON object. All keys which are not IRIs, compact IRIs, terms valid in the active context, or one of the following keywords MUST be ignored when processed:

If the node object contains the @context key, its value MUST be one of the following:

If the node object contains the @id key, its value MUST be an absolute IRI, a relative IRI, or a compact IRI (including blank node identifiers). See , , and for further discussion on @id values.

If the node object contains the @type key, its value MUST be either an absolute IRI, a relative IRI, a compact IRI (including blank node identifiers), a term defined in the active context expanding into an absolute IRI, or an array of any of these. See for further discussion on @type values.

If the node object contains the @graph key, its value MUST be a node object or an array of zero or more node objects. If the node object contains an @id keyword, its value is used as the label of a named graph. See for further discussion on @graph values. As a special case, if a JSON object contains no keys other than @graph and @context, and the JSON object is the root of the JSON-LD document, the JSON object is not treated as a node object; this is used as a way of defining node definitions that may not form a connected graph. This allows a context to be defined which is shared by all of the constituent node objects.

If the node object contains the @index key, its value MUST be a string. See section for further discussion on @index values.

Keys in a node object that are not keywords MUST expand to an absolute IRI using the active context. The values associated with these keys MUST be one of the following:


A term is a short-hand string that expands to an IRI or a blank node identifier.

A term MUST NOT equal any of the JSON-LD keywords.

To avoid forward-compatibility issues, a term SHOULD NOT start with an @ character as future versions of JSON-LD may introduce additional keywords. Furthermore, the use of empty terms ("") is discouraged as not all programming languages are able to handle empty property names.

See and for further discussion on mapping terms to IRIs.

Language Map

A language map is used to associate a language with a value in a way that allows easy programmatic access. A language map may be used as a term value within a node object if the term is defined with @container set to @language. The keys of a language map MUST be lowercase [[BCP47]] strings with an associated value that is any of the following types:

See for further discussion on language maps.

Index Map

An index map allows keys that have no semantic meaning, but should be preserved regardless, to be used in JSON-LD documents. An index map may be used as a term value within a node object if the term is defined with @container set to @index. The values of the members of an index map MUST be one of the following types:

See for further information on this topic.

Expanded Values

An expanded value is used to explicitly associate a type or a language with a value to create a typed value or a language-tagged string.

An expanded value MUST be a JSON object containing the @value key. It MAY also contain a @type, a @language, or an @index key but MUST NOT contain both a @type and a @language key at the same time. An expanded value MUST NOT contain keys other than @value, @type, @language, and @index. An expanded value that contains a @type key is called an expanded typed value. An expanded value that contains a @language key is called an expanded language-tagged string.

The value associated with the @value key MUST be either a string, number, true, false or null.

The value associated with the @type key MUST be a term, a compact IRI, an absolute IRI, a relative IRI, or null.

The value associated with the @language key MUST have the lexical form described in [[!BCP47]], or be null.

The value associated with the @index key MUST be a string.

See and for more information on expanded values.

List and Set Values

A list represents an ordered set of values. A set represents an unordered set of values. Unless otherwise specified, arrays are unordered in JSON-LD. As such, the @set keyword, when used in the body of a JSON-LD document, represents just syntactic sugar which is optimized away when processing the document. However, it is very helpful when used within the context of a document. Values of terms associated with a @set or @list container will always be represented in the form of an array when a document is processed - even if there is just a single value that would otherwise be optimized to a non-array form in compact document form. This simplifies post-processing of the data as the data is always in array form.

A list MUST be a JSON object that contains no other keys than @list, @context, and @index.

A set MUST be a JSON object that that contains no other keys than @set, @context, and @index. Please note that the @index key will be ignored, and thus be dropped, when being processed.

In both cases, the value associated with the keys @list and @set MUST be an array of any of the following:

See for further discussion on List and Set Values.

Context Definition

A context definition defines a local context in a node object.

A context definition MUST be a JSON object containing one or more key-value pairs. Keys MUST either be terms or @language or @vocab keywords.

If the context definition has a @language key, its value MUST have the lexical form described in [[!BCP47]] or be null.

If the context definition has a @vocab key, its value MUST have the lexical form of absolute IRI or be null.

Term values MUST be either a string, null, or an expanded term definition.

An expanded term definition is used to describe the mapping between a term and its expanded identifier, as well as other properties of the value associated with the term when it is used as key in a node object.

An expanded term definition SHOULD be a JSON object composed of zero or more keys from @id, @type, @language or @container. An expanded term definition SHOULD NOT contain any other keys.

If the term definition is not null, a compact IRI, or an absolute IRI and the active context does not have an @vocab mapping, the expanded term definition MUST include the @id key.

If the expanded term definition contains the @id keyword, its value MUST be null, an absolute IRI, a blank node identifier, a compact IRI, a term defined in the defining context definition or the active context, or an array composed of any of the previous allowed values except null.

If the expanded term definition contains the @type keyword, its value MUST be an absolute IRI, a compact IRI, a term defined in the defining context definition or the active context, null, or the one of the keywords @id or @vocab.

If the expanded term definition contains the @language keyword, its value MUST have the lexical form described in [[!BCP47]] or be null.

If the expanded term definition contains the @container keyword, its value MUST be either @list, @set, @language, @index, or be null. If the value is @language, when the term is used outside of the @context, the associated value MUST be a language map. If the value is @index, when the term is used outside of the @context, the associated value MUST be an index map.

Terms MUST NOT be used in a circular manner. That is, the definition of a term cannot depend on the definition of another term if that other term also depends on the first term.

See for further discussion on contexts.

Relationship to RDF

The RDF data model, as outlined in [[RDF-CONCEPTS]], is an abstract syntax for representing a directed graph of information. It is a subset of JSON-LD's data model with a few additional constraints. The differences between the two data models are:

Summarized these differences mean that JSON-LD is capable of serializing any RDF graph or dataset and most, but not all, JSON-LD documents can be transformed to RDF. A complete description of the algorithms to convert from RDF to JSON-LD and from JSON-LD to RDF is included in the JSON-LD Algorithms and API specification [[JSON-LD-API]].

Even though JSON-LD serializes RDF datasets, it can also be used as a RDF graph source. In that case, a consumer MUST only use the default graph and ignore all named graphs. This allows servers to expose data in, e.g., both Turtle and JSON-LD using content negotiation.

Publishers supporting both dataset and graph syntaxes have to ensure that the primary data is stored in the default graph to enable consumers that do not support datasets to process the information.

Relationship to Other Linked Data Formats

The JSON-LD markup examples below demonstrate how JSON-LD can be used to express semantic data marked up in other linked data formats such as Turtle, RDFa, Microformats, and Microdata. These sections are merely provided as evidence that JSON-LD is very flexible in what it can express across different Linked Data approaches.


The following are examples of converting RDF expressed in Turtle [[TURTLE-TR]] into JSON-LD.

Prefix definitions

The JSON-LD context has direct equivalents for the Turtle @prefix declaration:

JSON-LD has no equivalent for the Turtle @base declaration, but can use a prefix such as base to encode the information in the document.


Both Turtle and JSON-LD allow embedding, although Turtle only allows embedding of blank nodes.


Both JSON-LD and Turtle can represent sequential lists of values.


The following example describes three people with their respective names and homepages in RDFa [[RDFA-CORE]].

An example JSON-LD implementation using a single context is described below.


The following example uses a simple Microformats hCard example to express how Microformats [[MICROFORMATS]] are represented in JSON-LD.

The representation of the hCard expresses the Microformat terms in the context and uses them directly for the url and fn properties. Also note that the Microformat to JSON-LD processor has generated the proper URL type for http://tantek.com/.


The HTML Microdata [[MICRODATA]] example below expresses book information as a Microdata Work item.

Note that the JSON-LD representation of the Microdata information stays true to the desires of the Microdata community to avoid contexts and instead refer to items by their full IRI.

IANA Considerations

This section is included merely for standards community review and will be submitted to the Internet Engineering Steering Group if this specification becomes a W3C Recommendation.


Type name:
Subtype name:
Required parameters:
Optional parameters:

A whitespace-separated list of IRIs identifying specific constraints or conventions that apply to a JSON-LD document. A profile MUST NOT change the semantics of the resource representation when processed without profile knowledge, so that clients both with and without knowledge of a profiled resource can safely use the same representation. The profile parameter MAY also be used by clients to express their preferences in the content negotiation process. It is RECOMMENDED that profile IRIs are dereferenceable and provide useful documentation at that IRI. For more information and background please refer to [[PROFILE-LINK]].

This specification defines four values for the profile parameter. To request or specify Expanded JSON-LD document form, the IRI http://www.w3.org/ns/json-ld#expanded SHOULD be used. To request or specify Expanded, Flattened JSON-LD document form, the IRI http://www.w3.org/ns/json-ld#expanded-flattened SHOULD be used. To request or specify Compacted JSON-LD document form, the IRI http://www.w3.org/ns/json-ld#compacted SHOULD be used. To request or specify Compacted, Flattened JSON-LD document form, the IRI http://www.w3.org/ns/json-ld#compacted-flattened SHOULD be used. Please note that, according [[HTTP11]], the value of the profile parameter has to be enclosed in quotes (") because it contains special characters and, in some cases, whitespace.

Encoding considerations:
See RFC 6839, section 3.1.
Security considerations:
Since JSON-LD is intended to be a pure data exchange format for directed graphs, the serialization SHOULD NOT be passed through a code execution mechanism such as JavaScript's eval() function to be parsed.
JSON-LD contexts that are loaded from the Web over non-secure connections, such as HTTP, run the risk of modifying the JSON-LD active context in a way that could compromise security. It is advised that any application that depends on a remote context for mission critical purposes vet and cache the remote context before allowing the system to use it.
JSON-LD allows the substitution of long IRIs with short terms and the compression of multiple properties into a single property generator. Therefore, JSON-LD documents may expand enormously when processed and, in the worst case, the resulting data might consume all of the recipient's resources. Applications should treat any data with due skepticism.
Interoperability considerations:
Not Applicable
Published specification:
The JSON-LD specification.
Applications that use this media type:
Any programming environment that requires the exchange of directed graphs. Implementations of JSON-LD have been created for JavaScript, Python, Ruby, PHP, and C++.
Additional information:
Magic number(s):
Not Applicable
File extension(s):
Macintosh file type code(s):
Person & email address to contact for further information:
Manu Sporny <msporny@digitalbazaar.com>
Intended usage:
Restrictions on usage:
Manu Sporny, Dave Longley, Gregg Kellogg, Markus Lanthaler, Niklas Lindström
Change controller:

Fragment identifiers used with application/ld+json resources MAY identify a node in a JSON-LD graph expressed in the resource. This idiom, which is also used in RDF [[RDF-CONCEPTS]], gives a simple way to "mint" new, document-local IRIs to label nodes and therefore contributes considerably to the expressive power of JSON-LD.


The authors would like to extend a deep appreciation and the most sincere thanks to Mark Birbeck, who contributed foundational concepts to JSON-LD via his work on RDFj. JSON-LD uses a number of core concepts introduced in RDFj, such as the context as a mechanism to provide an environment for interpreting JSON data. Mark had also been very involved in the work on RDFa as well. RDFj built upon that work. JSON-LD exists because of the work and ideas he started nearly a decade ago in 2004.

A large amount of thanks goes out to the JSON-LD Community Group participants who worked through many of the technical issues on the mailing list and the weekly telecons - of special mention are François Daoust, Stéphane Corlosquet, Lin Clark, and Zdenko 'Denny' Vrandečić.

The work of David I. Lehn and Mike Johnson are appreciated for reviewing, and performing several early implementations of the specification. Thanks also to Ian Davis for this work on RDF/JSON.

Thanks to the following individuals, in order of their first name, for their input on the specification: Adrian Walker, Alexandre Passant, Andy Seaborne, Ben Adida, Blaine Cook, Bradley Allen, Brian Peterson, Bryan Thompson, Conal Tuohy, Dan Brickley, Danny Ayers, Daniel Leja, Dave Reynolds, David I. Lehn, David Wood, Dean Landolt, Ed Summers, elf Pavlik, Eric Prud'hommeaux, Erik Wilde, Fabian Christ, Jon A. Frost, Gavin Carothers, Glenn McDonald, Guus Schreiber, Henri Bergius, Jose María Alvarez Rodríguez, Ivan Herman, Jack Moffitt, Josh Mandel, KANZAKI Masahide, Kingsley Idehen, Kuno Woudt, Larry Garfield, Mark Baker, Mark MacGillivray, Marko Rodriguez, Melvin Carvalho, Nathan Rixham, Olivier Grisel, Paolo Ciccarese, Pat Hayes, Patrick Logan, Paul Kuykendall, Pelle Braendgaard, Peter Williams, Pierre-Antoine Champin, Richard Cyganiak, Roy T. Fielding, Sandro Hawke, Srecko Joksimovic, Stephane Fellah, Steve Harris, Ted Thibodeau Jr., Thomas Steiner, Tim Bray, Tom Morris, Tristan King, Sergio Fernández, Werner Wilms, and William Waites.