Copyright © 2026 the Linux Foundation
The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “NOT RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.
This document is licensed under The Apache License, Version 2.0.
Being able to express specific sequences of calls and articulate the dependencies between them to achieve a particular goal is desirable in the context of API descriptions. The aim of the Arazzo Specification is to provide a mechanism that can define sequences of calls and their dependencies to be woven together and expressed in the context of delivering a particular outcome or set of outcomes when dealing with API descriptions (such as OpenAPI descriptions).
The Arazzo Specification can articulate these workflows in a human-readable and machine-readable manner, thus improving the capability of API specifications to tell the story of the API in a manner that can improve the consuming developer experience.
A self-contained document (or set of documents) which defines or describes API workflows (specific sequence of calls to achieve a particular goal in the context of an API definition). An Arazzo Description uses and conforms to the Arazzo Specification, and MUST contain a valid Arazzo Specification version field (arazzo), an info field, a sourceDescriptions field with at least one defined Source Description, and there MUST be at least one Workflow defined in the workflows fixed field.
The Arazzo Specification is versioned using a major.minor.patch versioning scheme. The major.minor portion of the version string (for example 1.0) SHALL designate the Arazzo feature set. .patch versions address errors in, or provide clarifications to, this document, not the feature set. The patch version SHOULD NOT be considered by tooling, making no distinction between 1.0.0 and 1.0.1 for example.
An Arazzo Description that conforms to the Arazzo Specification is itself a JSON object, which may be represented either in JSON or YAML format.
All field names in the specification are case sensitive. This includes all fields that are used as keys in a map, except where explicitly noted that keys are case insensitive.
In order to preserve the ability to round-trip between YAML and JSON formats, YAML version 1.2 is RECOMMENDED along with some additional constraints:
It is RECOMMENDED that the entry Arazzo document be named: arazzo.json or arazzo.yaml.
An Arazzo Description MAY be made up of a single document or be divided into multiple, connected parts at the discretion of the author. If workflows from other documents are being referenced, they MUST be included as a Source Description Object. In a multi-document description, the document containing the Arazzo Specification Object is known as the entry Arazzo document.
Data types in the Arazzo Specification are based on the types supported by the JSON Schema Specification Draft 2020-12. Note that integer as a type is also supported and is defined as a JSON number without a fraction or exponent part.
As defined by the JSON Schema Validation vocabulary, data types can have an optional modifier property: format. Arazzo additionally supports the formats (similar to the OpenAPI specification) to provide fine detail for primitive data types.
The formats defined are:
format |
JSON Data Type | Comments |
|---|---|---|
int32 |
number | signed 32 bits |
int64 |
number | signed 64 bits (a.k.a long) |
float |
number | |
double |
number | |
password |
string | A hint to obscure the value. |
Each document in an Arazzo Description MUST be fully parsed in order to locate possible reference targets before attempting to resolve references.
This includes the parsing requirements of JSON Schema Specification Draft 2020-12, with appropriate modifications regarding base URIs as specified in Relative References in Arazzo Description URIs.
Reference targets include the Arazzo Object’s $self field (when present) and Source Description url fields.
Implementations MUST NOT treat a reference as unresolvable before completely parsing all documents provided to the implementation as possible parts of the Arazzo Description.
Fragmentary parsing occurs when an implementation parses only the specific part of a document being referenced, rather than parsing the complete document first. This practice is strongly discouraged and produces undefined behavior.
If only the referenced part of a document is parsed when resolving a reference, implementations may miss critical fields such as:
$self field, causing identity-based references to failurl fields, preventing proper resolution of referenced API descriptionsWhen fragmentary parsing causes references to resolve to unintended locations or fail to resolve entirely, the resulting behavior is undefined and implementation-defined. Different implementations may handle such cases inconsistently, breaking interoperability.
Implementations MUST parse entire documents before resolving references to ensure consistent, predictable behavior across different tools and environments.
To ensure interoperability, when referencing an Arazzo Description by URI, references MUST use the target document’s $self URI if the $self field is present in that document.
This means implementations MUST examine potential target Arazzo Descriptions (including those referenced via Source Description url fields or Workflow workflowId references) to check for matching $self values. When a Source Description Object’s url field is an absolute URI, the implementation MUST resolve it to an Arazzo Description by identity (matching the $self value) rather than by location alone.
Implementations MAY support referencing by other identifiers (such as retrieval URIs), but this behavior is not interoperable and relying on it is NOT RECOMMENDED.
URIs used as references within an Arazzo Description, including Source Description url, are resolved as identifiers, and described by this specification as URIs.
Unless specified otherwise, all fields that are URIs MAY be relative references as defined by RFC3986 Section 4.2.
Relative URI references are resolved using the appropriate base URI, which MUST be determined in accordance with RFC3986 Section 5.1.1 – 5.1.4.
The base URI for resolving relative references within an Arazzo Description is determined as follows:
$self field is present and is an absolute URI, the base URI is the $self URI (per RFC3986 Section 5.1.1: Base URI Embedded in Content).$self field is present and is a relative URI-reference, it MUST first be resolved against the next possible base URI source per RFC3986 Section 5.1.2 – 5.1.4 before being used as the base URI for resolving other relative references.$self field is not present, the base URI MUST be determined from the next possible base URI source per RFC3986 Section 5.1.2 – 5.1.4. The most common base URI source in this case is the retrieval URI of the Arazzo Description document (per RFC3986 Section 5.1.3), though other sources such as encapsulating entities (RFC3986 Section 5.1.2) or application-specific defaults (RFC3986 Section 5.1.4) MAY apply.For examples demonstrating base URI determination and reference resolution, see Appendix B: Examples of Base URI Determination and Reference Resolution.
If a URI contains a fragment identifier, then the fragment MUST be resolved per the fragment resolution mechanism of the referenced document.
For JSON or YAML documents, the fragment identifier SHOULD be interpreted as a JSON Pointer as per [RFC6901].
For JSON Schema objects used in workflow inputs and components.inputs, references via $ref MUST support both JSON Pointer fragments and plain-name fragments (defined by $anchor) as specified by JSON Schema Specification Draft 2020-12.
When an Arazzo Description references an OpenAPI or AsyncAPI description via a Source Description url field, any fragments in runtime expressions (such as in operationPath) use JSON Pointer syntax to identify specific operations or components within that description. The resolution of further references within the referenced OpenAPI or AsyncAPI description (including Schema Object $ref and $anchor keywords) MUST follow the rules and guidance laid out by the OpenAPI Specification and AsyncAPI Specification respectively.
Example:
sourceDescriptions:
- name: petstore
url: https://api.example.com/petstore.yaml
type: openapi
workflows:
- workflowId: example
steps:
- stepId: getPet
# Fragment '#/paths/~1pets/get' resolves via JSON Pointer
operationPath: '{$sourceDescriptions.petstore.url}#/paths/~1pets/get'
Relative references in CommonMark hyperlinks (such as those in description or summary fields) are resolved in their rendered context, which might differ from the context of the Arazzo Description.
API endpoints accessed during workflow execution are described by this specification as URLs (locations, not identifiers).
When Step Objects reference API operations via operationId or operationPath, the actual API endpoint URL is determined by the OpenAPI description’s Server Object, not by the Arazzo Description’s base URI.
Runtime expressions may reference API URLs via $url during workflow execution, but these are evaluated at execution time, not during document parsing.
In the following description, if a field is not explicitly REQUIRED or described with a MUST or SHALL, it can be considered OPTIONAL.
This is the root object of the Arazzo Description.
| Field Name | Type | Description |
|---|---|---|
| arazzo | string |
REQUIRED. This string MUST be the version number of the Arazzo Specification that the Arazzo Description uses. The arazzo field MUST be used by tooling to interpret the Arazzo Description. |
| $self | string |
A URI-reference for the Arazzo Description. This string MUST be in the form of a URI-reference as defined by RFC3986 Section 4.1. When present, this field provides the self-assigned URI of this Arazzo Description, which also serves as its base URI in accordance with RFC3986 Section 5.1.1 for resolving relative references within this document. The $self URI MUST NOT contain a fragment identifier. Arazzo Description documents can include a $self field to ensure portable, unambiguous reference resolution. |
| info | Info Object | REQUIRED. Provides metadata about the workflows contain within the Arazzo Description. The metadata MAY be used by tooling as required. |
| sourceDescriptions | [Source Description Object] | REQUIRED. A list of source descriptions (such as an OpenAPI description) this Arazzo Description SHALL apply to. The list MUST have at least one entry. |
| workflows | [Workflow Object] | REQUIRED. A list of workflows. The list MUST have at least one entry. |
| components | Components Object | An element to hold various schemas for the Arazzo Description. |
This object MAY be extended with Specification Extensions.
arazzo: 1.1.0
$self: https://api.example.com/workflows/pet-purchase.arazzo.yaml
info:
title: A pet purchasing workflow
summary: This Arazzo Description showcases the workflow for how to purchase a pet through a sequence of API calls
description: |
This Arazzo Description walks you through the workflow and steps of `searching` for, `selecting`, and `purchasing` an available pet.
version: 1.0.1
sourceDescriptions:
- name: petStoreDescription
url: https://github.com/swagger-api/swagger-petstore/blob/master/src/main/resources/openapi.yaml
type: openapi
- name: asyncOrderApiDescription
url: https://raw.githubusercontent.com/OAI/Arazzo-Specification/main/examples/1.1.0/pet-asyncapi.yaml
type: asyncapi
workflows:
- workflowId: loginUserAndRetrievePet
summary: Login User and then retrieve pets
description: This workflow lays out the steps to login a user and then retrieve pets
inputs:
type: object
properties:
username:
type: string
password:
type: string
orderCorrelationId:
type: string
steps:
- stepId: loginStep
description: This step demonstrates the user login step
operationId: $sourceDescriptions.petstoreDescription.loginUser
parameters:
# parameters to inject into the loginUser operation (parameter name must be resolvable at the referenced operation and the value is determined using {expression} syntax)
- name: username
in: query
value: $inputs.username
- name: password
in: query
value: $inputs.password
successCriteria:
# assertions to determine step was successful
- condition: $statusCode == 200
outputs:
# outputs from this step
tokenExpires: $response.header.X-Expires-After
rateLimit: $response.header.X-Rate-Limit
sessionToken: $response.body
- stepId: getPetStep
description: retrieve a pet by status from the GET pets endpoint
operationPath: '{$sourceDescriptions.petstoreDescription.url}#/paths/~1pet~1findByStatus/get'
parameters:
- name: status
in: query
value: 'available'
- name: Authorization
in: header
value: $steps.loginStep.outputs.sessionToken
successCriteria:
- condition: $statusCode == 200
onSuccess:
- name: 'noPetsAvailable'
type: "end"
criteria:
- condition: $response.body#/0 == null
outputs:
petId: $response.body#/0/id
- stepId: purchasePetStep
description: purchase a pet by posting an message on place-order channel
operationPath: $sourceDescriptions.asyncOrderApiDescription.placeOrder
action: send
parameters:
- name: orderCorrelationId
in: header
value: $inputs.orderCorrelationId
requestBody:
contentType: application/json
payload:
petId: $steps.getPetStep.outputs.petId
- stepId: confirmPetPurchaseStep
description: confirm the purchase of a pet
operationPath: $sourceDescriptions.asyncOrderApiDescription.confirmOrder
correlationId: $inputs.orderCorrelationId
timeout: 6000
action: receive
outputs:
orderId: $message.payload.orderId
outputs:
orderId: $steps.confirmPetPurchaseStep.outputs.orderId
The object provides metadata about API workflows defined in this Arazzo document. The metadata MAY be used by the clients if needed.
| Field Name | Type | Description |
|---|---|---|
| title | string |
REQUIRED. A human readable title of the Arazzo Description. |
| summary | string |
A short summary of the Arazzo Description. |
| description | string |
A description of the purpose of the workflows defined. CommonMark syntax MAY be used for rich text representation. |
| version | string |
REQUIRED. The version identifier of the Arazzo document (which is distinct from the Arazzo Specification version). |
This object MAY be extended with Specification Extensions.
title: A pet purchasing workflow
summary: This workflow showcases how to purchase a pet through a sequence of API calls
description: |
This workflow walks you through the steps of searching for, selecting, and purchasing an available pet.
version: 1.0.1
Describes a source description (such as an OpenAPI description) that will be referenced by one or more workflows described within an Arazzo Description.
An object storing a map between named description keys and location URLs to the source descriptions (such as an OpenAPI description) this Arazzo Description SHALL apply to. Each source location string MUST be in the form of a URI-reference as defined by [RFC3986] Section 4.1.
| Field Name | Type | Description |
|---|---|---|
| name | string |
REQUIRED. A unique name for the source description. Tools and libraries MAY use the name to uniquely identify a source description, therefore, it is RECOMMENDED to follow common programming naming conventions. SHOULD conform to the regular expression [A-Za-z0-9_\-]+. |
| url | string |
REQUIRED. A URL to a source description to be used by a workflow. If a relative reference is used, it MUST be in the form of a URI-reference as defined by [RFC3986] Section 4.2. |
| type | string |
The type of source description. Possible values are "openapi" or "asyncapi" or "arazzo". |
This object MAY be extended with Specification Extensions.
name: petStoreDescription
url: https://github.com/swagger-api/swagger-petstore/blob/master/src/main/resources/openapi.yaml
type: openapi
Describes the steps to be taken across one or more APIs to achieve an objective. The workflow object MAY define inputs needed in order to execute workflow steps, where the defined steps represent a call to an API operation or another workflow, and a set of outputs.
| Field Name | Type | Description |
|---|---|---|
| workflowId | string |
REQUIRED. Unique string to represent the workflow. The id MUST be unique amongst all workflows described in the Arazzo Description. The workflowId value is case-sensitive. Tools and libraries MAY use the workflowId to uniquely identify a workflow, therefore, it is RECOMMENDED to follow common programming naming conventions. SHOULD conform to the regular expression [A-Za-z0-9_\-]+. |
| summary | string |
A summary of the purpose or objective of the workflow. |
| description | string |
A description of the workflow. CommonMark syntax MAY be used for rich text representation. |
| inputs | JSON Schema |
A JSON Schema 2020-12 object representing the input parameters used by this workflow. |
| dependsOn | [string] |
A list of workflows that MUST be completed before this workflow can be processed. Each value provided MUST be a workflowId. If the workflow depended on is defined within the current Workflow Document, then specify the workflowId of the relevant local workflow. If the workflow is defined in a separate Arazzo Document then the workflow MUST be defined in the sourceDescriptions and the workflowId MUST be specified using a Runtime Expression (e.g., $sourceDescriptions.<name>.<workflowId>) to avoid ambiguity or potential clashes. |
| steps | [Step Object] | REQUIRED. An ordered list of steps where each step represents a call to an API operation or to another workflow. |
| successActions | [Success Action Object | Reusable Object] | A list of success actions that are applicable for all steps described under this workflow. These success actions can be overridden at the step level but cannot be removed there. If a Reusable Object is provided, it MUST link to success actions defined in the components/successActions of the current Arazzo document. The list MUST NOT include duplicate success actions. |
| failureActions | [Failure Action Object | Reusable Object] | A list of failure actions that are applicable for all steps described under this workflow. These failure actions can be overridden at the step level but cannot be removed there. If a Reusable Object is provided, it MUST link to failure actions defined in the components/failureActions of the current Arazzo document. The list MUST NOT include duplicate failure actions. |
| outputs | Map[string, {expression} | Selector Object ] |
A map between a friendly name and a dynamic output value defined using a Runtime Expression or Selector Object. The name MUST use keys that match the regular expression: ^[a-zA-Z0-9\.\-_]+$. |
| parameters | [Parameter Object | Reusable Object] | A list of parameters that are applicable for all steps described under this workflow. These parameters can be overridden at the step level but cannot be removed there. Each parameter MUST be passed to an operation or workflow as referenced by operationId, operationPath, or workflowId as specified within each step. If a Reusable Object is provided, it MUST link to a parameter defined in the components/parameters of the current Arazzo document. The list MUST NOT include duplicate parameters. |
This object MAY be extended with Specification Extensions.
workflowId: loginUser
summary: Login User
description: This workflow lays out the steps to login a user
inputs:
type: object
properties:
username:
type: string
password:
type: string
steps:
- stepId: loginStep
description: This step demonstrates the user login step
operationId: loginUser
parameters:
# parameters to inject into the loginUser operation (parameter name must be resolvable at the referenced operation and the value is determined using {expression} syntax)
- name: username
in: query
value: $inputs.username
- name: password
in: query
value: $inputs.password
successCriteria:
# assertions to determine step was successful
- condition: $statusCode == 200
outputs:
# outputs from this step
tokenExpires: $response.header.X-Expires-After
rateLimit: $response.header.X-Rate-Limit
outputs:
tokenExpires: $steps.loginStep.outputs.tokenExpires
Describes a single workflow step which MAY be a call to an API operation (OpenAPI Operation Object), (AysncAPI Operations Object) or another Workflow Object.
| Field Name | Type | Description |
|---|---|---|
| description | string |
A description of the step. CommonMark syntax MAY be used for rich text representation. |
| stepId | string |
REQUIRED. Unique string to represent the step. The stepId MUST be unique amongst all steps described in the workflow. The stepId value is case-sensitive. Tools and libraries MAY use the stepId to uniquely identify a workflow step, therefore, it is RECOMMENDED to follow common programming naming conventions. SHOULD conform to the regular expression [A-Za-z0-9_\-]+. |
| operationId | string |
The name of an existing, resolvable operation, as defined with a unique operationId and existing within one of the sourceDescriptions. The referenced operation will be invoked by this workflow step. If multiple (non arazzo type) sourceDescriptions are defined, then the operationId MUST be specified using a Runtime Expression (e.g., $sourceDescriptions.<name>.<operationId>) to avoid ambiguity or potential clashes. This field is mutually exclusive of the operationPath and workflowId fields respectively. |
| operationPath | string |
A reference to a Source Description Object combined with a JSON Pointer to reference an operation. This field is mutually exclusive of the operationId and workflowId fields respectively. The operation being referenced MUST be described within one of the sourceDescriptions descriptions. A Runtime Expression syntax MUST be used to identify the source description document. If the referenced operation has an operationId defined then the operationId SHOULD be preferred over the operationPath. |
| channelPath | string |
A reference to a Source Description Object combined with a JSON Pointer to reference an event channel. This field is mutually exclusive of the operationId and workflowId fields respectively. The operation being referenced MUST be described within one of the sourceDescriptions descriptions. A Runtime Expression syntax MUST be used to identify the source description document. If the referenced operation has an operationId defined then the operationId SHOULD be preferred over the channelPath. |
| workflowId | string |
The workflowId referencing an existing workflow within the Arazzo Description. If the referenced workflow is contained within an arazzo type sourceDescription, then the workflowId MUST be specified using a Runtime Expression (e.g., $sourceDescriptions.<name>.<workflowId>) to avoid ambiguity or potential clashes. The field is mutually exclusive of the operationId and operationPath fields respectively. |
| parameters | [Parameter Object | Reusable Object] | A list of parameters that MUST be passed to an operation or workflow as referenced by operationId, operationPath, or workflowId. If a parameter is already defined at the Workflow, the new definition will override it but can never remove it. If a Reusable Object is provided, it MUST link to a parameter defined in the components/parameters of the current Arazzo document. The list MUST NOT include duplicate parameters. |
| requestBody | Request Body Object | The request body to pass to an operation as referenced by operationId or operationPath. The requestBody is fully supported in HTTP methods where the HTTP 1.1 specification [RFC9110] Section 9.3 explicitly defines semantics for “content” like request bodies, such as within POST, PUT, and PATCH methods. For methods where the HTTP specification provides less clarity—such as GET, HEAD, and DELETE—the use of requestBody is permitted but does not have well-defined semantics. In these cases, its use SHOULD be avoided if possible. |
| successCriteria | [Criterion Object] | A list of assertions to determine the success of the step. Each assertion is described using a Criterion Object. All assertions MUST be satisfied for the step to be deemed successful. If successCriteria is provided, it MUST contain at least one Criterion Object. |
| onSuccess | [Success Action Object | Reusable Object] | An array of success action objects that specify what to do upon step success. If omitted, the next sequential step shall be executed as the default behavior. If multiple success actions have similar criteria, the first sequential action matching the criteria SHALL be the action executed. If a success action is already defined at the Workflow, the new definition will override it but can never remove it. If a Reusable Object is provided, it MUST link to a success action defined in the components of the current Arazzo document. The list MUST NOT include duplicate success actions. |
| onFailure | [Failure Action Object | Reusable Object] | An array of failure action objects that specify what to do upon step failure. If omitted, the default behavior is to break and return. If multiple failure actions have similar criteria, the first sequential action matching the criteria SHALL be the action executed. If a failure action is already defined at the Workflow, the new definition will override it but can never remove it. If a Reusable Object is provided, it MUST link to a failure action defined in the components of the current Arazzo document. The list MUST NOT include duplicate failure actions. |
| outputs | Map[string, {expression} | Selector Object] |
A map between a friendly name and a dynamic output value defined using a Runtime Expression or Selector Object. The name MUST use keys that match the regular expression: ^[a-zA-Z0-9\.\-_]+$. |
| timeout | integer |
The maximum number of milli-seconds to wait for the step to complete before aborting and failing the step. Consequently this will fail the workflow unless onFailure actions are defined. |
| correlationId | string |
A correlationId in AsyncAPI links a request with its response (or more broadly, to trace a single logical transaction across multiple asynchronous messages). Only applicable to asyncapi steps with action receive and has to be in-sync with correlationId defined in the AsyncAPI document. |
| action | string |
Describes the message flow intent. Indicates whether the step will send (publish) or receive (subscribe) to a channel in an AsyncAPI document. Only applicable for asyncapi steps. Possible values are "send" or "receive". |
| dependsOn | List[string] |
A list of steps that MUST be completed before this step can be executed. dependsOn only establishes a prerequisite relationship for the current step and does not trigger execution of the referenced steps. Each value provided MUST be a stepId. The stepId value is case-sensitive. If the step depended on is defined within the current workflow, specify the stepId directly (e.g., authStep). If the step is defined in a different workflow within the current Arazzo Document, reference it using $workflows.<workflowId>.steps.<stepId>. If the step is defined in a separate Arazzo Document, the workflow MUST be listed in sourceDescriptions and referenced using $sourceDescriptions.<name>.<workflowId>.steps.<stepId> to avoid ambiguity. If the step depends on the output of a non-blocking/asynchronous step, then it SHOULD use dependsOn and refer to the async step using one of these patterns. |
This object MAY be extended with Specification Extensions.
The dependsOn field at the step level is primarily intended to coordinate asynchronous operations.
For workflows containing only synchronous steps, the RECOMMENDED approach is to order steps sequentially in the steps array without using dependsOn. This provides the simplest and clearest execution model. Step-level dependsOn is typically unnecessary when all operations complete synchronously and execution follows a linear path.
When a step must wait for an asynchronous operation to complete before proceeding, dependsOn establishes a join point for in-flight async work. For example, a step that requests an order status SHOULD declare dependsOn on the step that receives order creation status from an async order placement, even if no explicit output reference exists. This ensures the async operation completes before the dependent step executes.
Tools MUST respect all declared dependsOn relationships. Tools MUST also treat runtime expression output references (e.g., $steps.stepId.outputs.field) as implicit dependencies and ensure the referenced step completes before the referencing step executes.
Tools supporting only sequential execution MUST execute steps in an order that satisfies both explicit (dependsOn) and implicit (output reference) dependencies.
Implementations SHOULD validate the following scenario:
When no dependsOn is used in a workflow (sequential execution model), implementations SHOULD produce an error if a step references another step’s outputs where the referenced step appears later in the steps array. This indicates a forward reference that cannot be satisfied in sequential execution.
For steps that reference AsyncAPI operations (via operationId or channelPath), tools will send or receive messages on the specified channel as defined by the referenced AsyncAPI description. For AsyncAPI send operations, the step completes immediately after the message is sent. Arazzo does not model broker acknowledgment or delivery confirmation. For AsyncAPI receive operations, step completion depends on message receipt as described below.
Because AsyncAPI channels MAY define multiple message types, and because message payloads MAY represent either success or failure conditions, authors SHOULD define successCriteria for AsyncAPI receive steps to explicitly evaluate the received message payload (for example, via $message.payload) and determine whether the step succeeded.
Authors MAY omit successCriteria only when both of the following conditions are met:
When successCriteria is omitted, receiving any message matching the correlationId (if specified) within the timeout period is considered a successful step completion.
If correlationId is specified, only messages matching the correlation identifier are considered. If no matching message is received within the timeout period, the step fails and triggers any defined onFailure actions.
A single step example:
stepId: loginStep
description: This step demonstrates the user login step
operationId: loginUser
parameters:
# parameters to inject into the loginUser operation (parameter name must be resolvable at the referenced operation and the value is determined using {expression} syntax)
- name: username
in: query
value: $inputs.username
- name: password
in: query
value: $inputs.password
successCriteria:
# assertions to determine step was successful
- condition: $statusCode == 200
outputs:
# outputs from this step
tokenExpires: $response.header.X-Expires-After
rateLimit: $response.header.X-Rate-Limit
A multiple step example:
steps:
- stepId: loginStep
description: This step demonstrates the user login step
operationId: loginUser
parameters:
# parameters to inject into the loginUser operation (parameter name must be resolvable at the referenced operation and the value is determined using {expression} syntax)
- name: username
in: query
value: $inputs.username
- name: password
in: query
value: $inputs.password
successCriteria:
# assertions to determine step was successful
- condition: $statusCode == 200
outputs:
# outputs from this step
tokenExpires: $response.header.X-Expires-After
rateLimit: $response.header.X-Rate-Limit
sessionToken: $response.body
- stepId: getPetStep
description: retrieve a pet by status from the GET pets endpoint
operationPath: '{$sourceDescriptions.petStoreDescription.url}#/paths/~1pet~1findByStatus/get'
parameters:
- name: status
in: query
value: 'available'
- name: Authorization
in: header
value: $steps.loginStep.outputs.sessionToken
successCriteria:
- condition: $statusCode == 200
outputs:
# outputs from this step
availablePets: $response.body
An async step example:
- stepId: placeOrder
description: This step demonstrates the action of sending a message payload to place an order
operationId: $sourceDescriptions.asyncOrderApi.placeOrder
action: send
parameters:
- name: requestId
in: header
value: $inputs.correlationId
requestBody:
payload:
productId: $inputs.productDetails.productId
quantity: $inputs.productDetails.quantity
- stepId: confirmOrder
description: This step demonstrates the action of receiving a message payload to confirm an order
operationId: $sourceDescriptions.asyncOrderApi.confirmOrder
correlationId: $inputs.correlationId
action: receive
dependsOn:
- placeOrder
timeout: 6000
outputs:
orderId: $message.payload.orderId
Describes a single step parameter. A unique parameter is defined by the combination of a name and in fields. There are several possible locations specified by the in field:
/items/{itemId}, the path parameter is itemId./items?id=###, the query parameter is id.content field (e.g., application/x-www-form-urlencoded). The querystring location cannot coexist with query parameters in the same operation per OpenAPI constraints.| Field Name | Type | Description |
|---|---|---|
| name | string |
REQUIRED. The name of the parameter. Parameter names are case sensitive. |
| in | string |
The location of the parameter. Possible values are "path", "query", "querystring", "header", or "cookie". When the step, success action, or failure action in context specifies a workflowId, then all parameters map to workflow inputs. In all other scenarios (e.g., a step specifies an operationId), the in field MUST be specified. |
| value | Any | {expression} | Selector Object | REQUIRED. The value to pass in the parameter. The value can be a constant, a Runtime Expression, or a Selector Object to be evaluated and passed to the referenced operation or workflow. For querystring parameters, the value MUST resolve to a string representing the complete query string (e.g., "key1=value1&key2=value2"). Runtime expressions can be embedded within the string value using {} notation. |
This object MAY be extended with Specification Extensions.
# Query Example
- name: username
in: query
value: $inputs.username
# Querystring Example (application/x-www-form-urlencoded)
- name: searchParams
in: querystring
value: "filter=active&sort=desc&limit=50"
# Querystring with Runtime Expressions (application/x-www-form-urlencoded)
- name: fullQuery
in: querystring
value: "category={$inputs.category}&minPrice={$inputs.minPrice}&inStock=true"
# Querystring Example (application/json)
- name: filterParams
in: querystring
value: '{"filter":"active","sort":"desc","limit":50}'
# Header Example
- name: X-Api-Key
in: header
value: $inputs.x-api-key
A single success action which describes an action to take upon success of a workflow step. There are two possible values for the type field:
workflowId or stepId)| Field Name | Type | Description |
|---|---|---|
| name | string |
REQUIRED. The name of the success action. Names are case sensitive. |
| type | string |
REQUIRED. The type of action to take. Possible values are "end" or "goto". |
| workflowId | string |
The workflowId referencing an existing workflow within the Arazzo Description to transfer to upon success of the step. This field is only relevant when the type field value is "goto". If the referenced workflow is contained within an arazzo type sourceDescription, then the workflowId MUST be specified using a Runtime Expression (e.g., $sourceDescriptions.<name>.<workflowId>) to avoid ambiguity or potential clashes. This field is mutually exclusive to stepId. |
| stepId | string |
The stepId to transfer to upon success of the step. This field is only relevant when the type field value is "goto". The referenced stepId MUST be within the current workflow. This field is mutually exclusive to workflowId. |
| parameters | [Parameter Object | Reusable Object] | A list of parameters that MUST be passed to a workflow as referenced by workflowId. If a Reusable Object is provided, it MUST link to a parameter defined in the components/parameters of the current Arazzo document. The list MUST NOT include duplicate parameters. The in field MUST NOT be used. |
| criteria | [Criterion Object] | A list of assertions to determine if this action SHALL be executed. Each assertion is described using a Criterion Object. All criteria assertions MUST be satisfied for the action to be executed. |
This object MAY be extended with Specification Extensions.
name: JoinWaitingList
type: goto
stepId: joinWaitingListStep
criteria:
# assertions to determine if this success action should be executed
- context: $response.body
condition: $[?count(@.pets) > 0]
type: jsonpath
A single failure action which describes an action to take upon failure of a workflow step. There are three possible values for the type field:
retryAfter and retryLimit fields. If a stepId or workflowId are specified, then the reference is executed and the context is returned, after which the current step is retried.workflowId or stepId)| Field Name | Type | Description |
|---|---|---|
| name | string |
REQUIRED. The name of the failure action. Names are case sensitive. |
| type | string |
REQUIRED. The type of action to take. Possible values are "end", "retry", or "goto". |
| workflowId | string |
The workflowId referencing an existing workflow within the Arazzo Description to transfer to upon failure of the step. This field is only relevant when the type field value is "goto" or "retry". If the referenced workflow is contained within an arazzo type sourceDescription, then the workflowId MUST be specified using a Runtime Expression (e.g., $sourceDescriptions.<name>.<workflowId>) to avoid ambiguity or potential clashes. This field is mutually exclusive to stepId. When used with "retry", context transfers back upon completion of the specified workflow. |
| stepId | string |
The stepId to transfer to upon failure of the step. This field is only relevant when the type field value is "goto" or "retry". The referenced stepId MUST be within the current workflow. This field is mutually exclusive to workflowId. When used with "retry", context transfers back upon completion of the specified step. |
| parameters | [Parameter Object | Reusable Object] | A list of parameters that MUST be passed to a workflow as referenced by workflowId. If a Reusable Object is provided, it MUST link to a parameter defined in the components/parameters of the current Arazzo document. The list MUST NOT include duplicate parameters. The in field MUST NOT be used. |
| retryAfter | number |
A non-negative decimal indicating the seconds to delay after the step failure before another attempt SHALL be made. Note: if an HTTP Retry-After response header was returned to a step from a targeted operation, then it SHOULD overrule this particular field value. This field only applies when the type field value is "retry". |
| retryLimit | integer |
A non-negative integer indicating how many attempts to retry the step MAY be attempted before failing the overall step. If not specified then a single retry SHALL be attempted. This field only applies when the type field value is "retry". The retryLimit MUST be exhausted prior to executing subsequent failure actions. |
| criteria | [Criterion Object] | A list of assertions to determine if this action SHALL be executed. Each assertion is described using a Criterion Object. |
This object MAY be extended with Specification Extensions.
name: retryStep
type: retry
retryAfter: 1
retryLimit: 5
criteria:
# assertions to determine if this action should be executed
- condition: $statusCode == 503
Holds a set of reusable objects for different aspects of the Arazzo Specification. All objects defined within the components object will have no effect on the Arazzo Description unless they are explicitly referenced from properties outside the components object.
Components are scoped to the Arazzo document they are defined in. For example, if a step defined in Arazzo document “A” references a workflow defined in Arazzo document “B”, the components in “A” are not considered when evaluating the workflow referenced in “B”.
| Field Name | Type | Description |
|---|---|---|
| inputs | Map[string, JSON Schema] |
An object to hold reusable JSON Schema objects to be referenced from workflow inputs. |
| parameters | Map[string, Parameter Object] |
An object to hold reusable Parameter Objects |
| successActions | Map[string, Success Action Object] |
An object to hold reusable Success Actions Objects. |
| failureActions | Map[string, Failure Action Object] |
An object to hold reusable Failure Actions Objects. |
This object MAY be extended with Specification Extensions.
All the fixed fields declared above are objects that MUST use keys that match the regular expression: ^[a-zA-Z0-9\.\-_]+$. The key is used to refer to the input or parameter in other parts of the Workflow Description.
Field Name Examples:
User
User_1
User_Name
user-name
my.org.User
components:
parameters:
storeId:
name: storeId
in: header
value: $inputs.x-store-id
inputs:
pagination:
type: object
properties:
page:
type: integer
format: int32
pageSize:
type: integer
format: int32
failureActions:
refreshToken:
name: refreshExpiredToken
type: retry
retryAfter: 1
retryLimit: 5
workflowId: refreshTokenWorkflowId
criteria:
# assertions to determine if this action should be executed
- condition: $statusCode == 401
"components": {
"parameters": {
"storeId": {
"name": "storeId",
"in": "header",
"value": "$inputs.x-store-id"
}
},
"inputs": {
"pagination": {
"type": "object",
"properties": {
"page": {
"type": "integer",
"format": "int32"
},
"pageSize": {
"type": "integer",
"format": "int32"
}
}
}
},
"failureActions": {
"refreshToken": {
"name": "refreshExpiredToken",
"type": "retry",
"retryAfter": 1,
"retryLimit": 5,
"workflowId": "refreshTokenWorkflowId",
"criteria": [
{
"condition": "$statusCode == 401"
}
]
}
}
}
A simple object to allow referencing of objects contained within the Components Object. It can be used from locations within steps or workflows in the Arazzo Description. Note - Input Objects MUST use standard JSON Schema referencing via the $ref keyword while all non JSON Schema objects use this object and its expression based referencing mechanism.
| Field Name | Type | Description |
|---|---|---|
| reference | {expression} |
REQUIRED. A Runtime Expression used to reference the desired object. |
| value | string |
Sets a value of the referenced parameter. This is only applicable for parameter object references. |
This object cannot be extended with additional properties and any properties added MUST be ignored.
reference: $components.successActions.notify
{
"reference": "$components.successActions.notify"
}
reference: $components.parameters.page
value: 1
{
"reference": "$components.parameters.page",
"value": 1
}
An object used to specify the context, conditions, and condition types that can be used to prove or satisfy assertions specified in Step Object successCriteria, Success Action Object criteria, and Failure Action Object criteria.
There are four flavors of conditions supported:
As part of a condition expression, you can use boolean, null, number, or string data types.
| Type | Literal value |
|---|---|
boolean |
true or false |
null |
null |
number |
Any number format supported in Data Types. |
string |
Strings MUST use single quotes (‘) around the string. To use a literal single quote, escape the literal single quote using an additional single quote (’'). |
| Operator | Description |
|---|---|
< |
Less than |
<= |
Less than or equal |
> |
Greater than |
>= |
Greater than or equal |
== |
Equal |
!= |
Not equal |
! |
Not |
&& |
And |
|| |
Or |
() |
Logical Grouping |
[] |
Index (0-based) |
. |
Property de-reference |
String comparisons MUST be case insensitive.
For simple conditions, Runtime Expressions can be used directly within the condition field:
# Direct usage in simple conditions
successCriteria:
- condition: $statusCode == 200
- condition: $response.body.count > $inputs.threshold
For regex, jsonpath, and xpath conditions, runtime expressions MUST be embedded within the condition string using {} curly braces. The runtime expressions are evaluated first, then substituted into the condition string before the expression is evaluated:
# Embedded expressions in JSONPath
successCriteria:
- context: $response.body
condition: '$[?(@.status == "{$inputs.expectedStatus}")]'
type: jsonpath
# Embedded expressions in XPath
successCriteria:
- context: $response.body
condition: '/root/items[price > {$steps.getPricing.outputs.minPrice}]'
type: xpath
# Embedded expressions in Regex
successCriteria:
- context: $response.body.status
condition: '^({$inputs.statusPattern})$'
type: regex
The evaluation order is as follows:
{} are evaluated and converted to stringstype (regex, jsonpath, xpath)The entire condition string MUST be quoted when it contains embedded expressions to ensure proper YAML parsing.
When runtime expressions are embedded in strings, type conversion follows the rules defined in Runtime Expressions.
A condition specified in a Criterion Object MUST evaluate to a pass (truthy) or fail (falsy) state. The evaluation semantics depend on the type of condition.
When type is simple or omitted, the condition MUST be an expression that combines Runtime Expressions, literals, and operators. The condition evaluates to:
A condition passes (truthy) when:
true.==, !=, <, >, <=, >=) evaluates to true.&&, ||) evaluates to true.A condition fails (falsy) when:
false.null.The following type conversion applies:
null only equals itself (null == null is true). Comparing null with any other value evaluates to false.Example:
# Pass if status code is 200
- condition: $statusCode == 200
# Pass if status code is 200 AND body contains data
- condition: $statusCode == 200 && $response.body.data != null
When type is regex, the condition MUST be a valid regular expression pattern, and context MUST be provided. The condition evaluates to:
context value.context value.If the context evaluates to null or undefined, the condition MUST evaluate to fail.
Example:
# Pass if status code starts with 2 (any 2xx code)
- context: $statusCode
condition: '^2\d{2}$'
type: regex
When type is jsonpath, the condition MUST be a valid JSONPath expression conforming to [RFC9535], and context MUST be provided.
JSONPath expressions return a NodesType (a nodelist, which is a sequence of zero or more nodes). The condition evaluates to:
If the context evaluates to null or undefined, or if the JSONPath expression is syntactically invalid, the condition MUST evaluate to fail.
JSONPath filter expressions (e.g., $[?count(@.pets) > 0]) that match nodes will return those nodes in the result nodelist. A filter with no matches returns an empty nodelist.
Example:
# Pass if response body contains at least one pet
- context: $response.body
condition: $[?count(@.pets) > 0]
type: jsonpath
# Pass if any pets array has elements
- context: $response.body
condition: $.pets[*]
type: jsonpath
When type is xpath, the condition MUST be a valid XPath expression conforming to the version specified (default: XML Path Language 3.1 or as specified using an Expression Type Object), and context MUST be provided.
XPath expressions can return different types: boolean, number, string, or node-set. The condition evaluates to:
A condition passes (truthy) when:
true (boolean)A condition fails (falsy) when:
false (boolean)If the context evaluates to null or undefined, or if the XPath expression is syntactically invalid, the condition MUST evaluate to fail.
Type conversion MUST follow the Effective Boolean Value (EBV) semantics defined by the XPath version being used. See Expression Type Object for version-specific semantics.
Example:
# Pass if count of pets is greater than 0 (returns boolean)
- context: $response.body
condition: count(/root/pets/*) > 0
type: xpath
# Pass if pets node exists (returns node-set)
- context: $response.body
condition: /root/pets
type: xpath
If a condition cannot be evaluated due to:
Then the condition MUST evaluate to fail, and implementations SHOULD log or report the error to aid debugging.
When multiple Criterion Objects are specified in successCriteria or criteria arrays, all conditions MUST evaluate to pass (truthy) for the overall criteria to be satisfied. This is equivalent to a logical AND operation across all criteria.
Example:
successCriteria:
# Both conditions must pass
- condition: $statusCode == 200
- context: $response.body
condition: $.data[*]
type: jsonpath
| Field Name | Type | Description |
|---|---|---|
| context | {expression} |
A Runtime Expression used to set the context for the condition to be applied on. If type is specified, then the context MUST be provided (e.g. $response.body would set the context that a JSONPath query expression could be applied to). |
| condition | string |
REQUIRED. The condition to apply. Conditions can be simple (e.g. $statusCode == 200 which applies an operator on a value obtained from a runtime expression), or a regex, or a JSONPath expression. For regex or JSONPath, the type and context MUST be specified. |
| type | string | Expression Type Object |
The type of condition to be applied. If specified, the options allowed are simple, regex, jsonpath or xpath. If omitted, then the condition is assumed to be simple, which at most combines literals, operators and Runtime Expressions. If jsonpath, then the expression MUST conform to JSONPath. If xpath the expression MUST conform to XML Path Language 3.1. Should other variants of JSONPath or XPath be required, then a Expression Type Object MUST be specified. |
This object MAY be extended with Specification Extensions.
A simple Condition example:
- condition: $statusCode == 200
A regex Condition example:
- context: $statusCode
condition: '^200$'
type: regex
A JSONPath Condition example:
- context: $response.body
condition: $[?count(@.pets) > 0]
type: jsonpath
An object used to describe the type and version of an expression used within a Criterion Object or Selector Object.
Defining this object gives the ability to utilize tooling compatible with older versions of either JSONPath or XPath.
| Field Name | Type | Description |
|---|---|---|
| type | string |
REQUIRED. The selector type. The options allowed are jsonpath, xpath, or jsonpointer. |
| version | string |
REQUIRED. A short hand string representing the version of the expression type being used. The allowed values for JSONPath are rfc9535 or draft-goessner-dispatch-jsonpath-00. The allowed values for XPath are xpath-30, xpath-20, or xpath-10. The allowed value for JSON Pointer is rfc6901. |
The supported expression selector types and versions are as follows:
| Type | Allowed Versions | Default |
|---|---|---|
jsonpath |
rfc9535, draft-goessner-dispatch-jsonpath-00 |
rfc9535 |
xpath |
xpath-31, xpath-30, xpath-20, xpath-10 |
xpath-31 |
jsonpointer |
rfc6901 (added for completeness) |
rfc6901 |
When used to specify a particular version of JSONPath or XPath, implementations MUST apply the semantics defined in that version’s specification. This includes:
For Criterion Object condition evaluation, the version-specific Effective Boolean Value (EBV) rules MUST be used when evaluating XPath expressions:
If this object is not defined, the default version for the selector type MUST be used.
This object MAY be extended with Specification Extensions.
A JSONPath example:
type: jsonpath
version: draft-goessner-dispatch-jsonpath-00
An XPath example:
type: xpath
version: xpath-30
An object which enables fine-grained traversal and precise data selection from structured data such as JSON or XML, using a defined selector syntax such as JSONPath or XPath.
| Field Name | Type | Description |
|---|---|---|
| context | {expression} | REQUIRED. A Runtime Expression which MUST evaluate to structured data (e.g., $response.body) and set the context for the selector to be applied on. |
| selector | string |
REQUIRED.A selector expression (e.g., $.items[0].id, /Envelope/Item) in the form of JSONPath expression, XPath expression, or JSON Pointer expression. |
| type | string | Expression Type Object |
REQUIRED. The selector expression type to use (e.g., jsonpath, xpath, or jsonpointer). If jsonpath, then the expression MUST conform to JSONPath. If xpath the expression MUST conform to XML Path Language 3.1. Should other variants of JSONPath or XPath be required, then a Expression Type Object MUST be specified. |
An output example:
outputs:
userEmail:
context: $response.body
selector: $.user.profile.email
type: jsonpath
A Step RequestBody example:
requestBody:
contentType: application/json
payload:
invoiceId:
context: $steps.fetchXml.outputs.invoiceXml
selector: /Invoice/Header/InvoiceNumber
type:
type: xpath
version: xpath-30
A single request body describing the Content-Type and request body content to be passed by a step to an operation.
| Field Name | Type | Description |
|---|---|---|
| contentType | string |
The Content-Type for the request content. If omitted then refer to Content-Type specified at the targeted operation to understand serialization requirements. |
| payload | Any | A value representing the request body payload. The value can be a literal value or can contain Runtime Expressions or Selector Objects which MUST be evaluated prior to calling the referenced operation. To represent examples of media types that cannot be naturally represented in JSON or YAML, use a string value to contain the example, escaping where necessary. |
| replacements | [Payload Replacement Object] | A list of locations and values to set within a payload. |
This object MAY be extended with Specification Extensions.
A JSON templated example:
contentType: application/json
payload: |
{
"petOrder": {
"petId": "{$inputs.pet_id}",
"couponCode": "{$inputs.coupon_code}",
"quantity": "{$inputs.quantity}",
"status": "placed",
"complete": false
}
}
A JSON object example:
contentType: application/json
payload:
petOrder:
petId: $inputs.pet_id
couponCode: $inputs.coupon_code
quantity: $inputs.quantity
status: placed
complete: false
A complete Runtime Expression example:
contentType: application/json
payload: $inputs.petOrderRequest
An XML templated example:
contentType: application/xml
payload: |
<petOrder>
<petId>{$inputs.pet_id}</petId>
<couponCode>{$inputs.coupon_code}</couponCode>
<quantity>{$inputs.quantity}</quantity>
<status>placed</status>
<complete>false</complete>
</petOrder>
A Form Data example:
contentType: application/x-www-form-urlencoded
payload:
client_id: $inputs.clientId
grant_type: $inputs.grantType
redirect_uri: $inputs.redirectUri
client_secret: $inputs.clientSecret
code: $steps.browser-authorize.outputs.code
scope: $inputs.scope
A Form Data String example:
contentType: application/x-www-form-urlencoded
payload: "client_id={$inputs.clientId}&grant_type={$inputs.grantType}&redirect_uri={$inputs.redirectUri}&client_secret={$inputs.clientSecret}&code{$steps.browser-authorize.outputs.code}&scope=$inputs.scope}"
Describes a location within a payload (e.g., a request body) and a value to set within the location.
| Field Name | Type | Description |
|---|---|---|
| target | string |
REQUIRED. A JSON Pointer, or XPath Expression, or JSONPath which MUST be resolved against the request body. Used to identify the location to inject the value. |
| targetSelectorType | string | Expression Type Object |
The selector expression type to use (e.g., jsonpath, xpath, or jsonpointer). If jsonpath, then the target expression MUST conform to JSONPath. If xpath the expression MUST conform to XML Path Language 3.1. Should other variants of JSONPath or XPath be required, then a Expression Type Object MUST be specified. |
| value | Any | {expression} | Selector Object | REQUIRED. The value set within the target location. The value can be a constant, a Runtime Expression, or Selector Objects to be evaluated and passed to the referenced operation or workflow. |
If targetSelectorType is omitted, then:
target MUST be interpreted as JSON Pointerif the payload is application/json.target MUST be interpreted as XPath Expression if the payload is application/xml or another XML-based media type.This object MAY be extended with Specification Extensions.
A Runtime Expression example:
target: /petId
value: $inputs.pet_id
A literal example:
target: /quantity
value: 10
A JSONPath example using an Expression Type Object:
target: $.items[?(@.sku=='ABC123')].quantity
targetSelectorType: jsonpath
value:
context: $steps.getInventory.outputs.payload
selector: $.newQuantity
type: jsonpath
An XPath example using older XPATH 3.0:
target: /Envelope/Header/CustomerId
targetSelectorType:
type: xpath
version: xpath-30
value:
context: $steps.fetchCustomerData.outputs.xml
selector: /CustomerInfo/Id
type:
type: xpath
version: xpath-30
A runtime expression allows values to be defined based on information that will be available within the HTTP message in an actual API call, or within objects serialized from the Arazzo document such as workflows or steps.
The runtime expression is defined by the following ABNF syntax:
; Top-level expression
expression = (
"$url" /
"$method" /
"$statusCode" /
"$request." source /
"$response." source /
"$message." source /
"$inputs." inputs-reference /
"$outputs." outputs-reference /
"$steps." steps-reference /
"$workflows." workflows-reference /
"$sourceDescriptions." source-reference /
"$components." components-reference /
"$self"
)
; Request/Response sources
source = ( header-reference / query-reference / path-reference / body-reference / payload-reference )
header-reference = "header." token
query-reference = "query." name
path-reference = "path." name
body-reference = "body" ["#" json-pointer ]
payload-reference = "payload" ["#" json-pointer ]
; Input/Output references
inputs-reference = input-name [ "#" json-pointer ]
outputs-reference = output-name [ "#" json-pointer ]
input-name = identifier
output-name = identifier
; Steps expressions
steps-reference = step-id ".outputs." output-name [ "#" json-pointer ]
step-id = identifier-strict
; Workflows expressions
workflows-reference = workflow-id "." workflow-field "." workflow-field-name [ "#" json-pointer ]
workflow-id = identifier-strict
workflow-field = "inputs" / "outputs"
workflow-field-name = identifier
; Source descriptions expressions
source-reference = source-name "." source-reference-id
source-name = identifier-strict
source-reference-id = 1*CHAR
; operationIds have no character restrictions in OpenAPI/AsyncAPI
; Resolution priority defined in spec text: (1) operationId/workflowId, (2) field names
; Components expressions
components-reference = component-type "." component-name
component-type = "parameters" / "successActions" / "failureActions"
component-name = identifier
; Identifier rules
identifier-strict = 1*( ALPHA / DIGIT / "-" / "_" )
; For step IDs, workflow IDs, and sourceDescription names (no dots)
; Matches [A-Za-z0-9_\-]+
identifier = 1*( ALPHA / DIGIT / "." / "-" / "_" )
; For component keys (dots allowed)
; Matches [a-zA-Z0-9\.\-_]+
name = *( CHAR )
; Allows unrestricted characters for query/path parameter names and field references
; JSON Pointer (RFC 6901)
json-pointer = *( "/" reference-token )
reference-token = *( unescaped / escaped )
unescaped = %x00-2E / %x30-7A / %x7C / %x7F-10FFFF
; Excludes / (%x2F), { (%x7B), } (%x7D), and ~ (%x7E)
escaped = "~" ( "0" / "1" )
; representing '~' and '/', respectively
; Token for header names (RFC 9110)
token = 1*tchar
tchar = "!" / "#" / "$" / "%" / "&" / "'" / "*" / "+" / "-" / "." /
"^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
; CHAR definition (RFC 7159, adapted to exclude { and })
CHAR = unescape / escape (
%x22 / ; " quotation mark U+0022
%x5C / ; \ reverse solidus U+005C
%x2F / ; / solidus U+002F
%x62 / ; b backspace U+0008
%x66 / ; f form feed U+000C
%x6E / ; n line feed U+000A
%x72 / ; r carriage return U+000D
%x74 / ; t tab U+0009
%x75 4HEXDIG ) ; uXXXX U+XXXX
escape = %x5C ; \
unescape = %x20-21 / %x23-5B / %x5D-7A / %x7C / %x7E-10FFFF
; Excludes { (%x7B) and } (%x7D) for unambiguous embedded expression parsing
; Expression strings
expression-string = *( literal-char / embedded-expression )
embedded-expression = "{" expression "}"
literal-char = %x00-7A / %x7C / %x7E-10FFFF
; Excludes { and } - simpler than CHAR for literal text
; Core ABNF rules (RFC 5234)
ALPHA = %x41-5A / %x61-7A ; A-Z / a-z
DIGIT = %x30-39 ; 0-9
HEXDIG = DIGIT / "A" / "B" / "C" / "D" / "E" / "F"
Here, json-pointer is taken from [RFC6901], CHAR from [RFC7159] Section 7 and token from [RFC7230] Section 3.2.6.
The name identifier is case-sensitive, whereas token is not.
| Source Location | example expression | notes |
|---|---|---|
| HTTP Method | $method |
The allowable values for the $method will be those for the HTTP operation. |
| Requested media type | $request.header.accept |
|
| Request parameter | $request.path.id |
Request parameters MUST be declared in the parameters section of the parent operation or they cannot be evaluated. This includes request headers. |
| Request body property | $request.body#/user/uuid |
In operations which accept payloads, references may be made to portions of the requestBody or the entire body. |
| Request URL | $url |
|
| Response value | $response.body#/status |
In operations which return payloads, references may be made to portions of the response body or the entire body. |
| Response array element | $response.body#/items/0/id |
Array elements can be accessed using numeric indices in JSON Pointer syntax. |
| Response header | $response.header.Server |
Single header values only are available. |
| Message header | $message.header.Server |
Single header values only are available. |
| Payload value | $message.payload#/status |
In operations which return payloads, references may be made to portions of the payload or the entire payload. |
| Self URI | $self |
References the canonical URI of the current Arazzo Description as defined by the $self field. |
| Workflow input | $inputs.username or $workflows.foo.inputs.username |
Single input values only are available. |
| Step output value | $steps.someStepId.outputs.pets |
In situations where the output named property return payloads, references may be made to portions of the response body (e.g., $steps.someStepId.outputs.pets#/0/id) or the entire body. |
| Workflow output value | $outputs.bar or $workflows.foo.outputs.bar |
In situations where the output named property return payloads, references may be made to portions of the response body (e.g., $workflows.foo.outputs.mappedResponse#/name) or the entire body. |
| Embedded expressions | https://{$inputs.host}/api/{$steps.create.outputs.id}/status |
Multiple runtime expressions can be embedded within a single string value by wrapping each in curly braces. |
| Source description reference | $sourceDescriptions.petstore.getPetById |
References an operationId or workflowId from the named source description. Resolution priority: (1) operationId/workflowId, (2) field names. |
| Source description field | $sourceDescriptions.petstore.url |
References a field from the Source Description Object. Resolved when no matching operationId/workflowId is found. |
| Components parameter | $components.parameters.foo |
Accesses a foo parameter defined within the Components Object. |
| Components action | $components.successActions.bar or $components.failureActions.baz |
Accesses a success or failure action defined within the Components Object. |
Runtime expressions preserve the type of the referenced value.
Expressions can be embedded into string values by surrounding the expression with {} curly braces. When a runtime expression is embedded in this manner, the following rules apply based on the value type:
Whether a value is stored as a string or parsed structure depends on its content type. JSON responses and inputs are typically parsed into structures, while XML and plain text are typically stored as strings. When embedding a parsed structure into a non-JSON payload format, the resulting JSON serialization may not match the target format’s expected structure.
When using $sourceDescriptions.<name>.<reference>, the <reference> portion is resolved with the following priority:
<reference> is first matched against operationIds. If the source description is an Arazzo document, <reference> is matched against workflowIds.<reference> is matched against field names of the Source Description Object (e.g., url,
type).Examples:
Given this source description:
sourceDescriptions:
- name: petstore
url: https://api.example.com/petstore.yaml
type: openapi
Given the above example source description and an OpenAPI description at that specified URL containing an operation with operationId: getPetById:
$sourceDescriptions.petstore.getPetById resolves to the operation with operationId getPetById (priority 1)$sourceDescriptions.petstore.url resolves to https://api.example.com/petstore.yaml (priority 2)$sourceDescriptions.petstore.type resolves to openapi (priority 2)If an operationId happens to conflict with a field name (e.g., an operation with operationId: url), the operationId takes precedence.
While the Arazzo Specification tries to accommodate most use cases, additional data can be added to extend the specification at certain points.
The extension properties are implemented as patterned fields that are always prefixed by "x-".
| Field Pattern | Type | Description |
|---|---|---|
| ^x- | Any | Allows extensions to the Arazzo Specification. The field name MUST begin with x-, for example, x-internal-id. Field names beginning x-oai-, x-oas-, and x-arazzo are reserved for uses defined by the OpenAPI Initiative. The value MAY be null, a primitive, an array or an object. |
The extensions may or may not be supported by the available tooling, but those may be extended as well to add requested support (if tools are internal or open-sourced).
The Arazzo Specification does not enforce a security mechanism. Security is left to the implementer, though TLS, specifically HTTPS may be recommended for exchanging sensitive workflows.
Arazzo Descriptions can be JSON or YAML values. As such, all security considerations defined in [RFC8259] and within YAML version 1.2 apply.
Arazzo Descriptions are frequently written by untrusted third parties, to be deployed on public Internet servers. Processing an Arazzo Description can cause both safe and unsafe operations to be performed on arbitrary network resources. It is the responsibility of the description consumer to ensure that the operations performed are not harmful.
The proposed MIME media types for the Arazzo Specification are described below.
The default (or general) MIME type for Arazzo documents (e.g. workflows) is defined as follows:
Media type name: application
Media subtype name: vnd.oai.workflows
Required parameters: N/A
Optional parameters: version (e.g. version=1.0.0 to indicate that the type of workflow conforms to version 1.0.0 of the Arazzo Specification).
Encoding considerations: Encoding considerations are identical to those specified for the application/json and application/yaml media types, respectively.
Security considerations: See security considerations above.
Interoperability considerations: N/A
Note: When using the application/vnd.oai.workflows media type the consumer should be prepared to receive YAML formatted content
The proposed MIME media type for Arazzo documents (e.g. workflows) that require a JSON-specific media type is defined as follows:
Media type name: application
Media subtype name: vnd.oai.workflows+json
Required parameters: N/A
Optional parameters: version (e.g. version=1.0.0 to indicate that the type of Arazzo document conforms to version 1.0.0 of the Arazzo Specification).
Encoding considerations: Encoding considerations are identical to those specified for the application/json media type.
Security considerations: See security considerations above.
Interoperability considerations: N/A
The proposed MIME media type for Arazzo documents (e.g. workflows) that require a YAML-specific media type is defined as follows:
Media type name: application
Media subtype name: vnd.oai.workflows+yaml
Required parameters: N/A
Optional parameters: version (e.g. version=1.0.0 to indicate that the type of Arazzo document conforms to version 1.0.0 of the Arazzo Specification).
Encoding considerations: Encoding considerations are identical to those specified for the application/yaml media type.
Security considerations: See security considerations above.
Interoperability considerations: N/A
| Version | Date | Notes |
|---|---|---|
| 1.1.0 | 2026-05-17 | Minor release of the Arazzo Specification 1.1.0 |
| 1.0.1 | 2025-01-16 | Patch release of the Arazzo Specification 1.0.1 |
| 1.0.0 | 2024-05-29 | First release of the Arazzo Specification |
This appendix provides concrete examples demonstrating how the $self field, Source Description URLs, and relative references work together across different deployment scenarios.
Assume the following Arazzo document is retrieved from file:///Users/dev/projects/workflows/purchase.arazzo.yaml:
arazzo: 1.1.0
$self: https://api.example.com/workflows/purchase.arazzo.yaml
info:
title: Pet Purchase Workflow
version: 1.0.0
sourceDescriptions:
- name: petstore
url: ../specs/petstore.yaml # Resolves to https://api.example.com/specs/petstore.yaml
type: openapi
The relative URL ../specs/petstore.yaml resolves against the $self base URI https://api.example.com/workflows/purchase.arazzo.yaml. The resolution algorithm per RFC3986 Section 5.2 removes the final path segment during resolution, producing https://api.example.com/specs/petstore.yaml, regardless of the retrieval URI.
If the same document does not define $self:
arazzo: 1.1.0
# No $self field
info:
title: Pet Purchase Workflow
version: 1.0.0
sourceDescriptions:
- name: petstore
url: ../specs/petstore.yaml
type: openapi
Retrieved from file:///Users/dev/projects/workflows/purchase.arazzo.yaml, the relative URL resolves to file:///Users/dev/projects/specs/petstore.yaml.
Per RFC3986 Section 5.1.2, the base URI can be provided by an encapsulating entity. For example, in a multipart/related response where an Arazzo Description is embedded:
Content-Type: multipart/related; boundary=example; type=application/vnd.oai.arazzo+json
--example
Content-Type: application/vnd.oai.arazzo+json
Content-Location: https://api.example.com/workflows/purchase.arazzo.json
{
"arazzo": "1.1.0",
"info": {...},
"sourceDescriptions": [
{
"name": "petstore",
"url": "../specs/petstore.json"
}
]
}
--example--
The Content-Location header provides the base URI (https://api.example.com/workflows/purchase.arazzo.json), so ../specs/petstore.json resolves to https://api.example.com/specs/petstore.json even without a $self field.
Per RFC3986 Section 5.1.4, applications may define default base URIs. For documents loaded without explicit retrieval URIs (e.g., from a database), implementations typically generate a unique base URI per document using a fixed prefix plus a unique identifier.
For example, a workflow orchestration platform might construct base URIs as https://workflows.example.com/{uuid} for each document:
arazzo: 1.1.0
# No $self field
# Loaded from database, assigned base URI: https://workflows.example.com/a7b3c4d5
info:
title: Pet Purchase Workflow
version: 1.0.0
sourceDescriptions:
- name: petstore
url: specs/petstore.yaml # Resolves using application default
If the application assigns base URI https://workflows.example.com/a7b3c4d5 to this document, then specs/petstore.yaml resolves to https://workflows.example.com/specs/petstore.yaml.
Note: While a base URI of https://workflows.example.com/ (with trailing slash) is technically valid per RFC3986, the final path component is an empty string. In practice, implementations typically assign unique identifiers as the final component to distinguish documents.
When $self is itself a relative URI-reference, it must be resolved before being used as a base URI:
arazzo: 1.1.0
$self: workflows/purchase.arazzo.yaml
info:
title: Pet Purchase Workflow
version: 1.0.0
sourceDescriptions:
- name: petstore
url: ../specs/petstore.yaml
Retrieved from https://api.example.com/v2/api-description.yaml:
$self relative reference workflows/purchase.arazzo.yaml against the base URI https://api.example.com/v2/api-description.yaml, which resolves to https://api.example.com/v2/workflows/purchase.arazzo.yaml per RFC3986 Section 5.2.url relative reference ../specs/petstore.yaml against the resolved $self base URI https://api.example.com/v2/workflows/purchase.arazzo.yaml, which resolves to https://api.example.com/v2/specs/petstore.yaml.An Arazzo Description may be retrieved from multiple locations but have a single canonical identity. Consider:
arazzo: 1.1.0
$self: https://workflows.example.com/canonical/purchase.arazzo.yaml
info:
title: Pet Purchase Workflow
version: 1.0.0
This document might be:
https://cdn.example.com/cache/abc123.yamlfile:///local/dev/purchase.arazzo.yamlmultipart/related responseIn all cases, references to this Arazzo Description MUST use https://workflows.example.com/canonical/purchase.arazzo.yaml (the $self value), not the retrieval location. This ensures that references remain stable even when the document is mirrored, cached, or moved.
Identity-based referencing via $self is particularly valuable in security-restricted environments. When deploying document sets behind firewalls or on air-gapped networks, implementations can scan for $self values to locate documents from a provided collection without making network requests that security policies might prevent. This enables reference resolution in environments where external network access is restricted or prohibited.