Retrieve All Agent Actions

Retrieves all agent actions in the currently authenticated Vault.

GET/api/{version}/services/ai/agent_actions

Headers

Name	Description
`Accept`	`application/json` (default) or `application/xml`

Query Parameters

Name	Description
`agent_name` optional	The name of an agent, for example, `promomats_document_chat__v`. Include to return the actions configured for this agent only. If omitted, retrieves all agent actions available to the authenticated user.
`available_by_user_name` optional	A Vault user name. Include to return the agent actions available to this user only. If omitted, retrieves all agent actions available to the authenticated user.
`object` conditional	The name of a Vault object. Required if `object_type` is specified. Include to return actions from agents configured on this object only. If omitted, retrieves actions from agents configured on all objects.
`object_type` optional	The name of a Vault object type. Include to return actions from agents configured on this object type only. If specified, must also set the `object` query parameter. If specified, must also set the `object` query parameter. If omitted, retrieves actions from agents configured on all object types.
`record_id` optional	A Vault object record ID. Include to return actions available on this record only. If specified, must also set the `object` query parameter. If omitted, retrieves actions available on all records.
`document_type` optional	The name of a document type. Include to return actions from agents configured on this document type only. If omitted, retrieves actions from agents configured on all document types.
`document_id` optional	A document ID, in the format `{ID}` or `{ID_major_minor}`. For example, `17` or `17_0_1`. Include to return actions available on the specified document only. If no version is specified, returns actions based on the latest version of the document. If `available_by_user_name` is also set, returns actions based on the latest version available to the specified user. If a document version is specified, returns actions based on the specified version. If `available_by_user_name` is also set, returns actions only if the specified user has access to the specified version.

Request

curl --location 'https://myvault.veevavault.com/api/v25.2/services/ai/agent_actions' \
--header 'Accept: application/json' \
--header 'Authorization: {SESSION_ID}' \
--header 'X-VaultAPI-ClientID: veeva-vault-api-reference-example'

Response

{
   "responseStatus": "SUCCESS",
   "agents": [
       {
           "agentName": "promomats_document_chat__v",
           "agentLabel": "Content Agent",
           "description": "This agent can be used to answer any user message, and provide assistance with questions or requests about the text and images in the current document the user is viewing, along with the Quick Check Results. It can also handle user messages that are unclear or ambiguous, including questions where the intent of the user message is not clear.",
           "actions": [
               {
                   "name": "document_text_questions__v",
                   "label": "Ask Questions",
                   "url": "/api/v25.2/services/ai/agent_action/promomats_document_chat__v/document_text_questions__v",
                   "description": "This action is used to answer any general question, request, or statement from the user, regarding the extracted document text, or the conversation history. Example requests include summarizing, extracting, analyzing, or reformatting content or prior messages. It can also handle messages that are not related to the data found in the document context view, helping to answer questions that are unclear or ambiguous, or messages where the user has not specified what they are trying to ascertain."
               },
               {
                   "name": "document_image_questions__v",
                   "label": "Analyze Images",
                   "url": "/api/v25.2/services/ai/agent_action/promomats_document_chat__v/document_image_questions__v",
                   "description": "This action is only used if the user message directly pertains to the **visual** elements of the document (such as images, logos, charts, diagrams, etc.), and cannot be answered using the text alone."
               },
               {
                   "name": "reviewer_summary__v",
                   "label": "Reviewer Summary",
                   "url": "/api/v25.2/services/ai/agent_action/promomats_document_chat__v/reviewer_summary__v",
                   "description": "This action must only be used if the user messages explicitly asks for a Reviewer Summary. It must not be used for any other kind of summary request."
               }
           ]
       },
       {
           "agentName": "quick_check__v",
           "agentLabel": "Quick Check",
           "description": "Performs automated compliance and quality checks on documents in regulated industries to support faster, smoother MLR review. Identifies issues in areas such as important safety information (ISI), boxed warnings, privacy and unsubscribe links, spelling, grammar, accessibility, and risky phrase assessment. Use only for structured document quality checks, not general Q&A.",
           "actions": [
               {
                   "name": "spelling_grammar__v",
                   "label": "Spelling and Grammar",
                   "url": "/api/v25.2/services/ai/agent_action/quick_check__v/spelling_grammar__v",
                   "description": "Performs a mechanical proofreading check on the document. Identifies spelling errors, typos, malformed technical terminology, and structural grammatical faults (syntax, morphology, and punctuation). Focuses strictly on linguistic accuracy and reading flow, leveraging domain-specific medical knowledge to identify technical errors while ignoring regulatory or scientific validity."
               },
               {
                   "name": "boxed_warning_assessment__v",
                   "label": "Boxed Warning Assessment",
                   "url": "/api/v25.2/services/ai/agent_action/quick_check__v/boxed_warning_assessment__v",
                   "description": "Analyzes document text to determine if a Boxed Warning is legally required based on its type (e.g., promotional, labeling). If required, it then uses full text and select page images to assess the presentation of the warning for visual and contextual compliance violations according to U.S. FDA regulations."
               }
           ]
       },
       {
           "agentName": "super_agent__sys",
           "agentLabel": "Super Agent",
           "description": "The super agent routes user requests to an appropriate agent action and coordinates multi-agent execution if necessary.",
           "actions": [
               {
                   "name": "route_action__sys",
                   "label": "Route Action",
                   "url": "/api/v25.2/services/ai/agent_action/super_agent__sys/route_action__sys",
                   "description": "This action analyzes the user message to match the user’s intent with an agent action to execute."
               }
           ]
       }
   ]
}

Response Details

On SUCCESS, the response contains details about each agent in the authenticated Vault, and each action the agent can perform.