Adobe Acrobat Knowledge Base (Preview)

Knowledge Base in Adobe Acrobat Studio allows users conversational access to knowledge that lives within large, curated document collections, across multi-turn Q&A, summarization, extraction, and search. Bring Knowledge Base capabilities into the applications and workflows where you work to connect knowledge stored in SharePoint and other supported repositories and power approvals, notifications, content creation, and business processes.

This connector is available in the following products and regions:

Service	Class	Regions
Copilot Studio	Premium	All Power Automate regions except the following: - US Government (GCC) - US Government (GCC High) - China Cloud operated by 21Vianet - US Department of Defense (DoD)
Logic Apps	Standard	All Logic Apps regions except the following: - Azure Government regions - Azure China regions - US Department of Defense (DoD)
Power Apps	Premium	All Power Apps regions except the following: - US Government (GCC) - US Government (GCC High) - China Cloud operated by 21Vianet - US Department of Defense (DoD)
Power Automate	Premium	All Power Automate regions except the following: - US Government (GCC) - US Government (GCC High) - China Cloud operated by 21Vianet - US Department of Defense (DoD)

Contact
Name	Adobe
URL	https://www.adobe.com

Connector Metadata
Publisher	ADOBE INC.
Website	https://business.adobe.com/products/acrobat-business/knowledge-base.html
Privacy policy	https://www.adobe.com/go/privacy_policy
Categories	Productivity

Knowledge Base in Acrobat

Publisher: Adobe, Inc.

Prerequisites

An Adobe account with access to Knowledge Base in Acrobat is required.

How to get credentials

Sign in with your Adobe ID when you create a connection. This connector authenticates with Adobe Identity Management System (IMS) using OAuth 2.0; you are redirected to the Adobe sign-in page and asked to consent to the requested permissions before the connection is created.

Get started with your connector

N/A

Known issues and limitations

N/A

Common errors and remedies

N/A

FAQ

N/A

Actions

Ask a question	Ask a natural-language question against a collection and get an answer with citations.
Create collection	Create a collection (a namespace that groups documents and controls access).
Delete attachment	Delete a single attachment by its UUID.
Delete collection	Delete a collection and all of its documents.
Delete documents	Delete one or more documents from a collection by their UUIDs.
Get attachment	Get a single attachment by its UUID.
Get attachment file	Download the file content of an attachment by its UUID.
Get collection	Get a single collection by its UUID.
Get document file	Get a single document's metadata and content by its UUID.
Get document status	Get the indexing status of a single document by its UUID.
List attachments	List attachments.
List collections	List the collections the user can access.
List documents	List the documents in a collection.
Search documents (paginated)	Search across documents in a collection with paginated ranked results.
Summarize documents	Generate a summary of one or more documents in a collection.
Update collection	Update all of an existing collection's properties.
Upload attachments	Upload one or more attachments.
Upload documents	Upload one or more documents to a collection for indexing.

Ask a question

Operation ID:: CreateQuestionAnswer

Ask a natural-language question against a collection and get an answer with citations.

Parameters

Name	Key	Required	Type	Description
Request Id	request_id		string	Unique identifier for this request, used for tracking in chat history.
Document Ids	document_ids		array of string	UUIDs of specific documents to restrict the search to. When empty, all documents in the specified collections are searched.
Collections	collections		array of string	Collection namespaces to search. At least one is required for COLLECTION source.
Inputs	inputs		string	The question or prompt to answer. Supports natural-language queries.
Format	format		string
Verbosity	style		string
Tone	tone		string
Reasoning	reasoning		string
Custom Instructions	custom_instructions		string	Free-text instructions to further customize the AI response behavior.
SourceOptions	Source Options		string
Inference Mode	inference_mode		string	AI model tier to use. Options: `STANDARD` (balanced speed and quality), `FAST_REASONING` (optimized for speed with reasoning), `ADVANCED` (highest quality), `LITE` (fast, lower cost — non-COLLECTION only), `NANO` (fastest, lowest cost, non-COLLECTION only). For `inference_type=AGENTIC` this is the tier of the agent that writes the answer; work it delegates may run on a faster tier chosen by the service.
Execution Strategy	inference_type		string	Execution strategy for question-answer inference. Orthogonal to InferenceMode (model tier): AGENTIC selects the multi-iteration agent loop, DEFAULT the classic single-inference RAG path. No missing coercion — unknown values must fail validation.
Key	key	True	string	Tag identifier. Alphanumeric characters and hyphens only.
Value	value	True	string	Tag value. Alphanumeric characters and hyphens only.
Attachment Ids	attachment_ids		array of uuid4	IDs of previously uploaded file attachments to include as visual context. Maximum 15 per request. Accepted input types: images (JPEG, PNG, GIF, WebP) and PDF. Other file types are rejected at upload time. In multi-turn conversations, the latest 5 LLM-generated files are automatically replayed as context; only images and PDFs are replayed due to LLM service infrastructure limitations.
File Generation Mode	file_generation		string
Connection Id	connection_id		string	ID of the Office 365 connection to use. Required only for OFFICE365 source. If omitted, the first matching connection is used.
Time Zone	time_zone		string	IANA time zone identifier for date-aware queries (e.g., `America/Los_Angeles`, `UTC`).
Enable Figures	enable_figures		boolean	When true, retrieved document figures (charts, images) are included as visual context. Up to 8 figures total (max 4 per document).
Store	store		boolean	When true, the conversation is persisted to chat history, enabling multi-turn follow-ups. Set to false for stateless, one-off queries.
AttributionField	Include		string	Fields from DocumentInfo eligible for inclusion in Source responses. Each value is a dot-notation path rooted at DocumentInfo. Every path must resolve to a field annotated with attribution_eligible = True.
BoundCoordinateFormat	bound_coordinate_format		string
Enabled	enabled		boolean	Enable progress events during streaming.
ResponseEventDetail	detail		string
Answer Updates	answer_updates		boolean	Stream the agentic answer as it is written, as frames carrying an answer_update (append, then complete or fail) instead of one final frame. The client must apply those operations. Ignored for non-streaming and non-agentic requests.
Exchange Id	exchange_id		string	The client's identifier of this exchange within the conversation. Stored with the exchange and returned in the chat history, so a client can reconcile its local turn with the record: the same question asked twice yields two exchanges, and only the id tells them apart.

Returns

QAInferenceResponse: QAInferenceResponse

Create collection

Operation ID:: CreateCollection

Create a collection (a namespace that groups documents and controls access).

Parameters

Name	Key	Required	Type	Description
Name	name	True	string	Display name for the collection. Required, max length 100 chars
Description	description		string	Optional description of the collection. Defaults to empty string, max length 500 chars
Is Public	is_public		boolean	Whether the collection is publicly accessible.
Is Frozen	is_frozen		boolean	Whether the collection is frozen and cannot be modified.
Read Users	read_users		array of string	List of user IDs with read access to the collection.
Write Users	write_users		array of string	List of user IDs with write access to the collection.
Read Groups	read_groups		array of string	List of group IDs (UUIDs) with read access to the collection.
Write Groups	write_groups		array of string	List of group IDs (UUIDs) with write access to the collection.
Metadata	metadata		object	Optional metadata associated with the collection. Max length 1000 chars

Returns

Complete collection details including metadata, permissions, and status.

NamespaceResponse: NamespaceResponse

Delete attachment

Operation ID:: DeleteAttachment

Delete a single attachment by its UUID.

Parameters

Name	Key	Required	Type	Description
Attachment ID	attachment_id	True	uuid	UUID of the attachment

Delete collection

Operation ID:: DeleteCollection

Delete a collection and all of its documents.

Parameters

Name	Key	Required	Type	Description
Collection ID	collection_id	True	uuid4	Unique identifier of the collection

Returns

Complete collection details including metadata, permissions, and status.

NamespaceResponse: NamespaceResponse

Delete documents

Operation ID:: DeleteDocuments

Delete one or more documents from a collection by their UUIDs.

Parameters

Name	Key	Required	Type	Description
Document Ids	document_ids		array of uuid4	List of document UUIDs to delete.
Collection Id	collection_id		string	UUID of the collection containing the documents.

Returns

DeleteDocumentResponse: DeleteDocumentResponse

Get attachment

Operation ID:: GetAttachment

Get a single attachment by its UUID.

Parameters

Name	Key	Required	Type	Description
Attachment ID	attachment_id	True	uuid	UUID of the attachment

Returns

AttachmentResponse: AttachmentResponse

Get attachment file

Operation ID:: GetAttachmentFile

Download the file content of an attachment by its UUID.

Parameters

Name	Key	Required	Type	Description
Attachment ID	attachment_id	True	uuid	UUID of the attachment

Get collection

Operation ID:: GetCollection

Get a single collection by its UUID.

Parameters

Name	Key	Required	Type	Description
Collection ID	collection_id	True	string	Unique identifier of the collection

Returns

Complete collection details including metadata, permissions, and status.

NamespaceResponse: NamespaceResponse

Get document file

Operation ID:: GetDocument

Get a single document's metadata and content by its UUID.

Parameters

Name	Key	Required	Type	Description
Document ID	document_id	True	uuid	UUID of the document
Artifact filename	filename		string	Specific stored artifact to retrieve. If omitted, the current document rendition is returned.

Returns

response: binary

Get document status

Operation ID:: GetDocumentStatus

Get the indexing status of a single document by its UUID.

Parameters

Name	Key	Required	Type	Description
Document ID	document_id	True	uuid	UUID of the document

Returns

Document Info: DocumentInfoApiResponse

List attachments

Operation ID:: ListAttachments

List attachments.

Returns

Response Listattachments: array of AttachmentResponse

List collections

Operation ID:: ListCollections

List the collections the user can access.

Parameters

Name Key Required Type Description

Page

page

integer

Page number (1-based)

Page size

page_size

integer

Number of items per page (1-100)

Order by

order_by

string

Field to sort results by.

Value	Sorts by
`name`	Collection display name (default)
`created_at`	Creation timestamp
`modified_at`	Last-modified timestamp

Ascending

asc

boolean

Sort in ascending order if true, descending if false.

Name contains

infix

string

Search term for collection name (case-insensitive partial match).

Filter

filter_constraint

string

Filter collections by ownership or scope.

Value	Returns
`created_by:me`	Collections you created
`shared_by:me`	Collections you shared with others
`shared_by:others`	Collections others shared with you
`scope:org`	All collections in your organization (org admin only)
`scope:system`	All collections across all organizations (system admin only)

Returns

CollectionPage: CollectionPage

List documents

Operation ID:: ListDocuments

List the documents in a collection.

Parameters

Name Key Required Type Description

Collection ID

collection_id

True

string

Unique identifier of the collection

Page

page

integer

Page number (1-based)

Page size

page_size

integer

Number of items per page (1-100)

Order by

order_by

string

Field to sort results by.

Value	Sorts by
`document_name`	Document filename (default)
`updated_at`	Last-modified timestamp
`upload_status`	Processing status
`uploaded_at`	Upload timestamp
`uploaded_by`	Uploader

Ascending

asc

boolean

Sort in ascending order if true, descending if false.

Name contains

infix

string

Search term for document name (case-insensitive partial match).

Source type

source_type

string

Restrict results to documents with this originating source. Documents without source metadata are excluded when this filter is set.

Returns

InfoPageApiResponse: InfoPageApiResponse

Search documents (paginated)

Operation ID:: SearchDocumentsPaginated

Search across documents in a collection with paginated ranked results.

Parameters

Name	Key	Required	Type	Description
Document Namespace	document_namespace	True	string	The collection name of the documents to search within.
Filter By	filter_by	True	string	Keyword search query to filter documents by content or metadata.
Document Ids	document_ids		array of string	List of specific document IDs to restrict the search to. Maximum of 1000 Document IDs.
Key	key	True	string	Tag identifier. Alphanumeric characters and hyphens only.
Value	value	True	string	Tag value. Alphanumeric characters and hyphens only.
Page Size	page_size		integer	The number of results to return per page. Non-paginated search always returns page_size of 10000.
Order By	order_by		string	The field to sort the results by.
Asc	asc		boolean	Sort order: True for ascending, False for descending.
Pagination Order	pagination_order		string	Determines whether to get the next or previous page of results.
Paginated Key	paginated_key		string	A key used for pagination.

Returns

PaginatedSearchApiResponse: PaginatedSearchApiResponse

Summarize documents

Operation ID:: SummarizeDocuments

Generate a summary of one or more documents in a collection.

Parameters

Name	Key	Required	Type	Description
Document Ids	document_ids	True	array of string	UUIDs of specific documents to summarize. Maximum 10 documents per request.
Document Namespace	document_namespace	True	string	Collection namespace containing the documents.

Returns

Response Summarizedocuments: array of SummarizationResponse

Update collection

Operation ID:: UpdateCollectionById

Update all of an existing collection's properties.

Parameters

Name	Key	Required	Type	Description
Collection ID	collection_id	True	uuid4	Unique identifier of the collection
Name	name	True	string	Display name for the collection. Required, max length 100 chars
Description	description		string	Optional description of the collection. Defaults to empty string, max length 500 chars
Is Public	is_public		boolean	Whether the collection is publicly accessible.
Is Frozen	is_frozen		boolean	Whether the collection is frozen and cannot be modified.
Read Users	read_users		array of string	List of user IDs with read access to the collection.
Write Users	write_users		array of string	List of user IDs with write access to the collection.
Read Groups	read_groups		array of string	List of group IDs (UUIDs) with read access to the collection.
Write Groups	write_groups		array of string	List of group IDs (UUIDs) with write access to the collection.
Metadata	metadata		object	Optional metadata associated with the collection. Max length 1000 chars
Namespace Id	namespace_id		string	UUID of the collection to update.

Returns

Complete collection details including metadata, permissions, and status.

NamespaceResponse: NamespaceResponse

Upload attachments

Operation ID:: UploadAttachments

Upload one or more attachments.

Parameters

Name	Key	Required	Type	Description
File to upload	documents	True	file	1–10 images or files. Images: image/jpeg, image/png, image/jpg, image/webp, image/gif - ≤10 MB each. Files: application/pdf - ≤8 MB each

Returns

API response for v2 attachment batch uploads.

AttachmentUploadsResponse: AttachmentUploadsResponse

Upload documents

Operation ID:: UploadDocuments

Upload one or more documents to a collection for indexing.

Parameters

Name Key Required Type Description

Collection ID

collection_id

True

uuid4

Unique identifier of the collection

Duplicate handling

option

string

Duplicate handling strategy.

Value	Behavior
`FILE_NAME`	Replace any existing document with the same filename
`FILE_CHECKSUM`	Replace the file only when the filename matches but the content hash (MD5) is different.
`SKIP_DUPLICATES`	Silently skip files that already exist (by checksum)
`REJECT_REQUEST_ON_DUPLICATE`	Return a 400 error if a duplicate is detected
(omitted)	Allow duplicate uploads

OCR mode

ocr

string

OCR processing mode for PDF and image documents. Image uploads always require OCR.

Value	Behavior
`AUTO`	Apply OCR when needed; image uploads are always processed (default)
`FORCE`	Always rasterize and OCR-process every page
`NONE`	Skip OCR for non-image documents; image uploads are still processed
`REJECT_BAD`	Apply OCR automatically and reject pages that fail quality checks

Hidden-sheet extraction

hidden_sheets

string

Controls hidden sheet extraction mode for Excel documents.

Value	Behavior
`ALL`	extract all sheets, including hidden and very hidden sheets (default)
`NONE`	extract only visible sheets
`HIDDEN`	extract visible and hidden sheets
`VERY_HIDDEN`	extract visible and very hidden sheets

File to upload

documents

True

file

One or more files to upload. Supported formats: PDF, DOC/DOCX, DOT/DOTX, TXT, RTF, XLS/XLSM/XLSX, PPT/PPTX, ODP, ODT, CSV, EML, JSON, MD, HTML, JPEG, PNG, TIFF, GIF, WebP, and BMP.

Document tags

document_tags

string

JSON array of key-value tags to attach to all uploaded documents. Each tag must have a key (1-127 chars, lowercase alphanumeric and hyphens) and a value (1-256 chars).

HTML source links

html_links

string

JSON object mapping HTML filenames to their original source URLs. Only applies to uploaded HTML files. The key is the filename of an HTML file in the upload, and the value is the original URL it was saved from.

Returns

Response Uploaddocuments: array of UploadDocumentResponse

Definitions

AnswerOutcome

One answer's outcome: none, complete, or incomplete with the code of what cut it.

Name	Path	Type	Description
Status	status	string	Whether an answer exists and whether its generation finished.
Reason	reason	string	The interruption code when incomplete, the failure code when none; never set when complete.

AnswerUpdate

One operation on the answer a streaming client shows: append, complete or fail.

complete is incomplete when reason is set; fail carries a failure code and never withdraws the shown text.

Name	Path	Type	Description
Operation	operation	string
Reason	reason	string	Terminal only: the interruption code on complete (e.g. response_truncated_deadline), the failure code on fail (e.g. budget_exhausted, save_failed).
Persistence	persistence	Persistence	Whether an exchange is on record; `unknown` means the commit may have happened.

AttachmentErrorCode

Error codes for failed attachment uploads.

AttachmentErrorCode: string

AttachmentResponse

Name	Path	Type	Description
Attachment Id	attachment_id	string
Attachment Name	attachment_name	string
Attachment Mime Type	attachment_mime_type	string	Image MIME type (e.g. image/png)
Attachment Checksum	attachment_checksum	string	MD5 checksum of the image bytes
Attachment Size	attachment_size	integer	Size of the image in bytes
Attachment Status	attachment_status	string

AttachmentUploadError

Machine-readable failure details for a single attachment upload.

Name	Path	Type	Description
AttachmentErrorCode	code	AttachmentErrorCode	Error codes for failed attachment uploads.
Message	message	string
Size	size	integer
Mime Type	mime_type	string

AttachmentUploadItem

Unified per-file result in the v2 attachment upload response.

Name	Path	Type	Description
Filename	filename	string
AttachmentUploadItemStatus	status	AttachmentUploadItemStatus	Per-file status for v2 attachment batch upload responses.
Status Code	status_code	integer
AttachmentResponse	attachment	AttachmentResponse
AttachmentUploadError	error	AttachmentUploadError	Machine-readable failure details for a single attachment upload.

AttachmentUploadItemStatus

Per-file status for v2 attachment batch upload responses.

AttachmentUploadItemStatus: string

AttachmentUploadsResponse

API response for v2 attachment batch uploads.

Name	Path	Type	Description
Uploads	uploads	array of AttachmentUploadItem

CollectionPage

Name	Path	Type	Description
Page	page	integer	Current page number (1-based).
Size	size	integer	Number of items per page.
Total	total	integer	Total number of collections matching the query.
Order By	order_by	string	Field used to sort results.
Asc	asc	boolean	Whether results are sorted in ascending order.
Pages	pages	array of Namespace	List of collections on the current page.

DeleteDocumentResponse

Name	Path	Type	Description
Details	details	array of DocumentDeletionStatus	Per-document deletion results.

DocumentDeletionStatus

Name	Path	Type	Description
Document Id	document_id	string	UUID of the document.
Status	status	Status
Message	message	string	Human-readable message describing the outcome.

DocumentInfoApiResponse

Name Path Type Description

Document Id

document_id

string

UUID of the document.

Document Name

document_name

string

Display name of the document (typically the original filename).

Document Path

document_path

string

Storage path of the document within the system.

Document Namespace

document_namespace

string

Collection namespace this document belongs to.

Document Namespace Id

document_namespace_id

string

UUID of the collection the document belongs to.

Document Description

document_description

string

User-provided description of the document.

Uploaded By

uploaded_by

string

Email address of the person who uploaded the document. Null when no uploader email is available.

Upload Status

upload_status

string

Current upload/processing status.

Value	Meaning
`UPLOADED`	File received, awaiting processing
`IN_PROGRESS`	Processing has started
`CONVERTING`	Converting to indexable format
`INDEXING`	Building search index
`COMPLETE`	Ready for queries
`FAILED`	Processing failed

Checksum

checksum

string

SHA-256 checksum of the document file for integrity verification.

Mime Type

mime_type

string

MIME type of the original source document.

Rendition Mime Type

rendition_mime_type

string

MIME type of the client rendition; external sources use text/uri-list.

Source Type

source_type

string

Originating source of this document: SHAREPOINT, SHAREPOINT_SITE, GOOGLE_DRIVE, or UPLOAD. null for documents without source metadata.

DocumentStatus

status

DocumentStatus

Uploaded At

uploaded_at

date-time

Timestamp when the document was uploaded (ISO 8601 format).

Updated At

updated_at

date-time

Timestamp of the last update to the document record, if any.

Is Deleted

is_deleted

boolean

Whether the document has been soft-deleted.

DocumentMetadata

document_metadata

DocumentMetadata

Name	Path	Type
Permissions	permissions
Document Tags	document_tags	array of DocumentTag
Source Metadata	source_metadata
IngestionMetadata	ingestion_metadata	IngestionMetadata

DocumentStatus

Name	Path	Type	Description
Status	status	string
Error Code	error_code	string

DocumentTag

Name	Path	Type	Description
Key	key	string	Tag identifier. Alphanumeric characters and hyphens only.
Value	value	string	Tag value. Alphanumeric characters and hyphens only.

InfoPageApiResponse

Name	Path	Type	Description
Page	page	integer	Current page number (1-based).
Size	size	integer	Number of items per page.
Total	total	integer	Total number of documents matching the query.
Order By	order_by	string	Field used to sort results.
Asc	asc	boolean	Whether results are sorted in ascending order.
Pages	pages	array of DocumentInfoApiResponse	List of documents on the current page.

IngestionMetadata

Name	Path	Type
Confidence Score	confidence_score	number
Ocr	ocr	boolean
Bad Characters	bad_characters	boolean
Found Hidden Sheets	found_hidden_sheets	boolean
Found Very Hidden Sheets	found_very_hidden_sheets	boolean
Found Protected Hidden Sheets	found_protected_hidden_sheets	boolean
Enabled Hidden Sheets	enabled_hidden_sheets	boolean
Enabled Very Hidden Sheets	enabled_very_hidden_sheets	boolean
Document Created Date	document_created_date	date-time

Namespace

Name	Path	Type	Description
Namespace	namespace	string	Unique identifier for the collection. Defaults to UUID. Use it as an external ID.
Name	name	string	Optional display name for the collection. Required if 'namespace' is not provided.
Description	description	string	Human-readable description of the collection.
Is Frozen	is_frozen	boolean	Whether the collection is frozen and cannot be modified.
Created By	created_by	string	Email or ID of the collection creator.
Created By User Id	created_by_user_id	string	User ID of the collection creator.
Is Public	is_public	boolean	Whether the collection is publicly accessible.
Read Users	read_users	array of string	List of user IDs with read access.
Write Users	write_users	array of string	List of user IDs with write access.
Read Groups	read_groups	array of string	List of group IDs with read access.
Write Groups	write_groups	array of string	List of group IDs with write access.
Metadata	metadata	object	Additional metadata associated with the collection.
Version	version	string	API version of the collection.
Namespace Id	namespace_id	string	Unique identifier for the collection.
Created At	created_at	date-time	Timestamp when the collection was created.
Modified At	modified_at	date-time	Timestamp when the collection was last modified.
Is Deleted	is_deleted	boolean	Whether the collection has been soft-deleted.
Is Active	is_active	boolean	Whether the collection is currently active.
Doc Count	doc_count	integer	Document count for the namespace. Expose as a top level property for namespace, from metadata.doc_count

NamespaceResponse

Complete collection details including metadata, permissions, and status.

Name	Path	Type	Description
Namespace Id	namespace_id	string	Unique identifier for the collection.
Namespace	namespace	string	URL-safe identifier used in API calls.
Name	name	string	Display name of the collection.
Description	description	string	Human-readable description of the collection.
Is Frozen	is_frozen	boolean	Whether the collection is frozen and cannot be modified.
Created By	created_by	string	User ID of the collection creator.
Is Public	is_public	boolean	Whether the collection is publicly accessible.
Read Users	read_users	array of string	List of user IDs with read access.
Write Users	write_users	array of string	List of user IDs with write access.
Read Groups	read_groups	array of string	List of group IDs with read access.
Write Groups	write_groups	array of string	List of group IDs with write access.
Rejected Write Shares	rejected_write_shares	array of string	Non entitled write share users
Metadata	metadata	object	Additional metadata associated with the collection.
Version	version	string	API version for this response.
Created At	created_at	date-time	Timestamp when the collection was created.
Modified At	modified_at	date-time	Timestamp when the collection was last modified.
Is Deleted	is_deleted	boolean	Whether the collection has been soft-deleted.
Is Active	is_active	boolean	Whether the collection is currently active.
Doc Count	doc_count	integer	Number of documents in the collection.
Owner	owner	string	Email of the collection owner. Currently the user who created the collection.

PaginatedSearchApiResponse

Name	Path	Type	Description
Totaldocuments	totalDocuments	integer	Total number of documents matching the search criteria.
Documentinfo	documentInfo	array of DocumentInfoApiResponse	List of documents in the current page of results.
Paginated Key	paginated_key	string	Opaque pagination token. Pass this value in subsequent requests to retrieve the next or previous page.

Persistence

Whether an exchange is on record; unknown means the commit may have happened.

Persistence: string

QAInferenceResponse

Name	Path	Type	Description
Generated Text	generated_text	string	The AI-generated answer in Markdown format.
Source	source	object	Source documents referenced in the answer, keyed by numeric citation markers (e.g., `1`, `2`). Citations appear in the generated text as `[1]`, `[2]`, etc.
Stream Complete	stream_complete	boolean	Indicates whether the streaming response has finished. Always true for non-streaming responses.
Source Options	source_options	array of SourceOptions	The answer source types that were used to generate this response.
Attachments	attachments	array of uuid4	IDs of LLM-generated file attachments (e.g., CSV, XLSX, JSON, PDF, images) that accompany the answer. Supported output types include text formats, spreadsheets, documents, and images. In multi-turn conversations, the latest 5 generated files are automatically replayed to the model; only images and PDFs are replayed as LLM input due to LLM service infrastructure limitations.
Reasoning Summary	reasoning_summary	string	A summary of the AI reasoning process, included when reasoning mode is enabled.
ResponseEvent	response_event	ResponseEvent
AnswerUpdate	answer_update	AnswerUpdate	One operation on the answer a streaming client shows: append, complete or fail. `complete` is incomplete when `reason` is set; `fail` carries a failure code and never withdraws the shown text.
AnswerOutcome	answer_outcome	AnswerOutcome	One answer's outcome: none, complete, or incomplete with the code of what cut it.

ResponseEvent

Name	Path	Type	Description
Type	type	string	Dot-namespaced event identifier following the pattern `response.status.{phase}.{state}`. Lifecycle events: `response.created`, `response.in_progress`, `response.error`. Status events: `context.loading` (loading conversation history for follow-up questions), `question.analyzing` (understanding the question), `documents.searching` (searching collection for relevant documents), `documents.found` (retrieval complete, documents identified), `sources.linking` (mapping answer evidence to source documents), `figures.analyzing` (analyzing document figures and charts), `context.preparing` (assembling document context for the model), `answer.generating` (model is generating the answer), `answer.streaming` (first answer token received, streaming in progress).
Message	message	string	User-facing progress text describing the current processing step.
Meta	meta	object	Structured metadata for the event. Common fields: `collection_names` (list of collection display names), `doc_count` (number of documents found), `question` (the analyzed question text), `document_names` (names of specific requested documents), `is_follow_up` (true for multi-turn follow-up questions).
Created At	created_at	string	ISO 8601 UTC timestamp of when this event was produced.

SourceOptions

SourceOptions: string

Status

Status: string

SummarizationResponse

Name	Path	Type	Description
Summary	summary	string	The AI-generated document summary in Markdown format.
Source	source	object	Source documents used to produce the summary, keyed by numeric citation marker.

UploadDocumentResponse

Response returned for each uploaded document.

Name	Path	Type	Description
User Id	user_id	string	ID of the user who uploaded the document.
Document Id	document_id	string	Unique identifier assigned to the uploaded document.
Document Name	document_name	string	Original filename of the uploaded document.
Document Namespace	document_namespace	string	Namespace (collection identifier) the document belongs to.
Document Namespace Id	document_namespace_id	string	UUID of the collection the document belongs to.
Status	status	string	Processing status of the uploaded document.
Content Type	content_type	string	MIME type of the uploaded document.

binary

This is the basic data type 'binary'.