스키마 | D.Hub 사용자 매뉴얼

📄️ AccessControlPolicy

Shared request and response schema. Request: Access-control policy identified by the composite `resource_id` and `subject_id` fields. It has no single `id`; row filters and column masks must use the structured fields rather than raw SQL. Response: Access-control policy identified by the composite `resource_id` and `subject_id` fields. It has no single `id`; row filters and column masks use the structured response fields.

📄️ ActorCallNodeConfig

Settings that control Actor Call Node. The required field is `actor`.

📄️ ActorCreate

Payload used to create Actor. Required fields are `name`, `description`, `function`.

📄️ ActorDuplicate

Actor duplicate (save-as / clone) request. Mirrors ``ToolDuplicate``: executable definition and confirmation policy come from the source actor; only identity/placement metadata is overridable.

📄️ ActorResponse

Data returned for Actor. Required fields are `id`, `name`, `category`, `description`, `function`, `item`.

📄️ ActorUpdate

Payload used to update Actor. Optional fields include `name`, `alias`, `description`, `function`, `confirmation_policy`, and 4 more.

📄️ AdminPasswordReset

Admin password reset request.

📄️ AgentActorConfig

Actor configuration for an agent node.

📄️ AgentCompaction

Per-agent context compaction policy.

📄️ AgentCreate

Payload used to create Agent. Required fields are `name`, `description`.

📄️ AgentNodeConfig

Settings that control Agent Node. Required fields are `instructions`, `model_provider`, `model`.

📄️ AgentPromptRef

Reusable prompt reference with a `prompt_id` and optional variable bindings. An agent may include multiple references; their list order is preserved.

📄️ AgentReferenceResponse

Agent location that references a first-class agent resource.

📄️ AgentResourceDeletePreviewRequest

Resources to check before a single or bulk delete.

📄️ AgentResourceDeletePreviewResponse

Delete preflight results in request order.

📄️ AgentResourceDeletePreviewTarget

One resource whose Agent references must be checked before deletion.

📄️ AgentResourceRef

Stable reference to a first-class resource used by an agent node.

📄️ AgentResourceReferencesResponse

Read-only delete preflight for resources referenced by agents.

📄️ AgentResponse

Data returned for Agent. Required fields are `id`, `name`, `category`, `description`, `item`.

📄️ AgentSessionSourceRef

Canonical reference to an owner-scoped Agent conversation.

📄️ AgentToolConfig

Tool configuration for an agent node.

📄️ AgentUpdate

Payload used to update Agent. Optional fields include `name`, `alias`, `description`, `input_mode`, `mode`, and 10 more.

📄️ AggSpec

Inline aggregation definition. `agg` is required; `expr` may be omitted only for `count`, and `percentile` must be between 0 and 1 when `agg` is `percentile`.

📄️ AggregateConfig

Config for the ``aggregate`` op: group-by + allowlisted aggregate functions. An empty ``group_by`` aggregates the whole input to a single row. ``input`` is a port key.

📄️ AggregateExpr

One aggregation output: an allowlisted function over a column, named by ``alias``. ``count`` may omit ``column`` (a row count). Raw SQL / arbitrary expressions are intentionally not representable.

📄️ AggregateTransform

The ``aggregate`` member of the ``PipelineTransform`` union.

📄️ Announcement

Announcement banner entry.

📄️ AnnouncementCreate

Request to create an announcement.

📄️ AnnouncementUpdate

Partial update for an announcement.

📄️ ArtifactTemplate

Listing entry for an artifact scaffold (no ZIP body).

📄️ ArtifactVersion

Uploaded artifact version with its owning resource, storage version identifier, byte size, digest, modification time, and latest-version status.

📄️ AssistantActiveModel

Active assistant model selected for the organization. If the selected model is unavailable, the API returns no active model.

📄️ AssistantActiveModelUpdate

Request that selects the organization-wide active assistant model by `model_id`.

📄️ AssistantLlmModelCreate

Request that registers an assistant model with its provider, model identifier, endpoint, and optional credential.

📄️ AssistantLlmModelResponse

Assistant-model configuration returned by the API, including whether a credential is stored without returning its value.

📄️ AssistantLlmModelRuntime

Assistant-model connection details returned to an authorized service caller, including the provider, model, endpoint, and credential needed to invoke the configured model.

📄️ AssistantLlmModelTestRequest

Assistant-model connection test request. `base_url` is required; omit `api_key` and provide `model_id` to reuse that model's stored credential.

📄️ AssistantLlmModelUpdate

Request that changes mutable assistant-model provider, model, endpoint, label, or credential fields.

📄️ AssistantProvidersResponse

Provider identifiers available when configuring an assistant model.

📄️ AssistantSessionSourceRef

Canonical reference to an owner-scoped Assistant conversation.

📄️ Batch

Batch execution resource.

📄️ BatchResponse

Batch start response — Batch representation + deprecated `item` mirror.

📄️ BatchState

Batch execution state.

📄️ Body_auth-upload_avatar

Multipart form data containing the avatar image to upload.

📄️ Body_collections-upload_to_collection

Multipart form data containing files to add to a collection.

📄️ Body_datasets-inference_schema_from_files

Multipart form data containing files whose columns and types are used for schema inference.

📄️ Body_datasets-update_object

Multipart form data containing the replacement file for an object in a dataset.

📄️ Body_datasets-upload_to_dataset

Multipart form data containing files to append to a dataset.

📄️ Body_imports-import_scenario

Multipart form data containing a scenario archive and optional resource or column name overrides.

📄️ Body_models-upload_model_artifact

Multipart form data containing a model artifact version to upload.

📄️ Body_oidc-introspect

Form data used to inspect whether an OIDC token is active.

📄️ Body_oidc-revoke

Form data identifying an OIDC token to revoke.

📄️ Body_oidc-token

Form data for authorization-code or refresh-token exchange.

📄️ Body_ontology-upload_to_entity_table

Multipart form data containing files to load into an entity table.

📄️ Body_ontology-upload_to_relation_table

Multipart form data containing files to load into a relation table.

📄️ Body_prompts-upload_prompt_document

Multipart form data containing a document and metadata for the prompt created from it.

📄️ Body_settings-upload_branding_image

Multipart form data containing the branding image to upload.

📄️ BrandingImages

Branding image URLs.

📄️ BrandingSettings

Settings that control Branding. Optional fields include `app_name`, `page_title`, `login_slogan`, `copyright_holder`, `copyright_start_year`, and 3 more.

📄️ BrandingSettingsUpdate

Partial update for branding settings.

📄️ CaptureCommit

Request used to save a web-session storage state as a secret, keyed by the target site and optional account, in an optional collection.

📄️ CaptureCreate

Request that starts a web-session capture, optionally scoped to a collection.

📄️ Checkpoint

Step execution checkpoint.

📄️ CleanupStatusResponse

Asynchronous resource-cleanup status with offsets, timing, resource identity, and an optional failure description.

📄️ CodeCreate

Payload used to create Code. Required fields are `name`, `type`.

📄️ CodeNodeConfig

Settings that control Code Node. The required field is `source`.

📄️ CodeResponse

Data returned for Code. Required fields are `id`, `name`, `category`, `type`, `item`.

📄️ CodeUpdate

Payload used to update Code. Optional fields include `name`, `alias`, `type`, `collection_id`, `folder_id`, and 3 more.

📄️ CodeUsage

A single pipeline whose step graph references a code.

📄️ CodeUsagesResponse

Reverse-lookup of pipelines referencing a code (mirrors DatasetUsagesResponse).

📄️ CollectionCreate

Payload used to create Collection. The required field is `name`.

📄️ CollectionNameAvailabilityResponse

Data returned for Collection Name Availability. Required fields are `scope`, `collection_id`, `name`, `available`.

📄️ CollectionResponse

Data returned for Collection. Required fields are `id`, `name`, `category`, `item`.

📄️ CollectionUpdate

Partial update for a collection. `name` must match the current name; rename by deleting and recreating the collection.

📄️ Column

Semantic column definition. Capability flags control filtering, sorting, grouping, and catalog visibility; `values` is exhaustive, `samples` is illustrative, and dictionary-backed columns suppress both value lists from semantic catalog output.

📄️ ColumnDictionary

Value-linking dictionary for a semantic column. Exactly one of `builtin` or `external` must be set: `builtin` contains an inline canonical-value-to-synonyms map, while `external` names a shared dictionary.

📄️ ColumnMask

Column masking or derivation rule. Access policies and pipeline inputs accept schema-preserving masks with no `source` or `expression`; views may reference backing columns and use every `ColumnMaskType` value.

📄️ ColumnMaskType

Unified column mask / derivation vocabulary. Schema-preserving value transforms are valid in every consumer. The derivation / aggregation entries (``expr``, ``sum``, ``count``, ``avg``, ``min``, ``max``) change row/column shape and are only valid in View — AccessControlPolicy and PipelineData declare their allowed subset locally and reject the rest.

📄️ ConnectorBuiltinTemplateResponse

Built-in connector template metadata exposed for connector creation UI.

📄️ ConnectorCreate

Payload used to create Connector. Required fields are `name`, `type`.

📄️ ConnectorCredentialInput

Request-only connector credential value.

📄️ ConnectorQueryRequest

Ad-hoc request through an existing connector. ``rest`` connectors only for now: the request is handed to the connector's verb functions — the connector script owns auth/transport specifics (e.g. the SAP BW ``$batch`` read tunnel). ``GET`` is always available; write verbs additionally require the connector to opt in via ``options.allow_console_writes``.

📄️ ConnectorQueryResponse

Connector query result. Successful responses include the returned body and any extracted rows; connector failures return `success=false` with `error` instead of raising an HTTP error.

📄️ ConnectorResponse

Data returned for Connector. Required fields are `id`, `name`, `category`, `type`, `item`.

📄️ ConnectorSummaryResponse

Connector list/selector response without secret-bearing fields.

📄️ ConnectorTestRequest

Connection test request for a draft or existing connector.

📄️ ConnectorTestResponse

Connection test result.

📄️ ConnectorTypeInfoResponse

Connector type info plus built-in templates for that type.

📄️ ConnectorUpdate

Payload used to update Connector. Optional fields include `name`, `alias`, `type`, `driver`, `collection_id`, and 7 more.

📄️ ConversationSourceSelection

Stable conversation range selected for a report source.

📄️ ConversationVariable

Conversation variable definition.

📄️ CypherQuery

Graph query text with optional parameters and target database name.

📄️ DashboardCreate

Payload used to create Dashboard. The required field is `name`.

📄️ DashboardResponse

Data returned for Dashboard. Required fields are `id`, `name`, `category`, `item`.

📄️ DashboardUpdate

Payload used to update Dashboard. Optional fields include `name`, `alias`, `description`, `widgets`, `collection_id`, and 4 more.

📄️ DashboardWidget

Data accepted by Create Dashboard. Required fields are `config`, `layout`.

📄️ DataSinkState

Current sink state entries returned for a dataset table.

📄️ DatasetCreate

Payload used to create Dataset. Required fields are `name`, `type`, `collection_id`.

📄️ DatasetObjectEntry

Single file entry of an object-type dataset (Delta metadata row).

📄️ DatasetResponse

Data returned for Dataset. Required fields are `id`, `name`, `category`, `type`, `item`.

📄️ DatasetUpdate

Partial update for a dataset. Do not include `name`; `collection_id` may move the dataset to another collection but cannot be cleared.

📄️ DatasetUsageEntity

Single ontology resource that backs its data from a dataset.

📄️ DatasetUsagesResponse

Reverse-lookup of ontology resources backed by a dataset.

📄️ DecoratorBinding

Configuration-based decorator binding that supplies or overrides decorator arguments without editing code. When kind and name match, the configured binding takes precedence; `decorator` must name a supported decorator.

📄️ DerivedRef

Derived-metric dependency identified by name, with an optional expression alias and comparison offset window.

📄️ DhubAssetSourceRef

D.Hub asset locator identified by source, asset type, and asset identifier. Access is checked with the current user's credentials when the report input is resolved, before the report run is submitted. `subresource_type` and `subresource_id` must be provided together; `version` is optional.

📄️ DictionaryCandidate

One nearest surface in the dictionary partition. ``name`` is the canonical item name (code), ``surface`` the matched surface form, ``score`` the cosine.

📄️ DictionaryItem

Canonical dictionary value with a display label and alternative surface forms used for value matching.

📄️ DictionaryItemListResponse

Page of dictionary items with the total item count, current page, and page size.

📄️ DictionaryItemsReplace

Atomic full replacement body — PUT /semantic/dictionaries/{id}/items.

📄️ DictionaryItemsReplaceResponse

Bounded response for a potentially million-row atomic replacement.

📄️ DictionaryItemsUpsert

Bulk item upsert body — POST /semantic/dictionaries/{id}/items.

📄️ DictionarySearchRequest

Preview request for value-linking search — POST /semantic/dictionaries/{id}/search.

📄️ DictionarySearchResponse

Dictionary search result containing the requested value and its nearest candidate values.

📄️ DirectorySettings

Settings that control Directory. Optional fields include `member_discovery`.

📄️ DirectorySettingsUpdate

`PUT /settings/directory` request body (admin-only).

📄️ DocumentCreate

Document creation request. `collection_id` is not accepted from clients — it is server-derived from the parent Knowledge by the create route.

📄️ DocumentResponse

Data returned for Document. Required fields are `id`, `name`, `category`, `item`.

📄️ DocumentUpdate

Document update request. No `collection_id` — a document's scope follows the parent Knowledge and cannot be changed independently.

📄️ DriftEntry

Schema-drift entry identifying the affected resource and table, missing columns, type mismatches, and first detection time.

📄️ DriftMarker

Detected schema-drift summary with the affected fields and detection time.

📄️ EndNodeConfig

Settings that control End Node. Optional fields include `outputs`.

📄️ EntityTypeCreate

Payload used to create Entity Type. Required fields are `name`, `collection_id`.

📄️ EntityTypeResponse

Data returned for Entity Type. Required fields are `id`, `name`, `category`, `item`.

📄️ EntityTypeUpdate

Partial update for an entity type. `name` must match the current name, and schema changes must be additive.

📄️ ErrorHandlerNodeConfig

Settings that control Error Handler Node. The required field is `target_node_id`.

📄️ Filter

Reusable named filter composed of one or more query clauses.

📄️ FilterClause

A compile-side filter clause (same shape as the query Condition, but author-defined). Used by a named ``Filter``, and as a filtered-aggregate clause on ``AggSpec``/``Metric``.

📄️ FolderCreate

Payload used to create Folder. Required fields are `name`, `collection_id`.

📄️ FolderResponse

Data returned for Folder. Required fields are `id`, `name`, `category`, `path`, `item`.

📄️ FolderUpdate

Payload used to update Folder. Optional fields include `name`, `alias`, `folder_id`, `tags`, `description`.

📄️ FooterLink

Footer link entry.

📄️ FunctionAction

Shared request and response schema. Request: Python-only reaction action accepted with executable code and input values. The return value is ignored, so the action is used for side effects. Response: Python-only reaction action returned with executable code and input values. The return value is ignored, so the action is used for side effects.

📄️ FunctionCondition

Cron-triggered Python condition. On each scheduled occurrence, the condition runs `code` with `params` and proceeds to the reaction actions only when the code returns true.

📄️ FunctionCreate

Payload used to create Function. Required fields are `name`, `script`.

📄️ FunctionMeta

Metadata for a single @function-decorated callable in a Function script.

📄️ FunctionResponse

Data returned for Function. Required fields are `id`, `name`, `category`, `item`.

📄️ FunctionUpdate

Payload used to update Function. Optional fields include `name`, `alias`, `collection_id`, `tags`, `description`, and 1 more.

📄️ GenerateMockDataPreviewResponse

Dry-run preview of generated sample rows — nothing is written.

📄️ GenerateMockDataRequest

Request body for schema-driven sample (mock) data generation.

📄️ GenerateMockDataResponse

Result of a sample data generation write.

📄️ GrafanaAlert

One alert inside the webhook payload's ``alerts[]``.

📄️ GrafanaWebhookPayload

Top-level Grafana Unified Alerting webhook body.

📄️ GroupCreate

Request model for creating a group (local only).

📄️ GroupMember

Response model for group member listing.

📄️ GroupResponse

Data returned for Group. Required fields are `name`, `item`.

📄️ GroupUpdate

Request model for partially updating a group.

📄️ HTTPValidationError

Validation failure response containing field-level error details.

📄️ HumanInputNodeConfig

Settings that control Human Input Node. The required field is `prompt`.

📄️ IdPGroupInfo

IdP group info for registration workflow.

📄️ IdPUserInfo

IdP user info for registration workflow.

📄️ IfElseNodeConfig

Settings that control If Else Node. The required field is `condition`.

📄️ ImageType

Branding image type.

📄️ InputField

User input field definition.

📄️ JoinColumn

A projected output column of a ``join``. ``input`` is a port key — ``left_input`` or ``right_input`` — and ``name`` a column on that input; ``alias`` optionally renames it in the output.

📄️ JoinCondition

Single binary condition in a Join.on list. AND'd within a Join.

📄️ JoinConfig

Two-input equi-join configuration. `left_input` and `right_input` are pipeline port keys, `keys` contains at least one equality pair, and `columns` explicitly selects the output. Other join predicates are not accepted.

📄️ JoinExpr

Join condition operand. Exactly one of `column` / `expression` set.

📄️ JoinKey

One equi-join column pair (``left`` column == ``right`` column).

📄️ JoinKind

Allowed Join Kind values. Values: `inner`, `left`, `right`, `full`.

📄️ JoinOperator

Allowed Join Operator values. Values: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`.

📄️ JoinTransform

The ``join`` member of the ``PipelineTransform`` union.

📄️ KnowledgeCreate

Payload used to create Knowledge. The required field is `name`.

📄️ KnowledgeResponse

Data returned for Knowledge. Required fields are `id`, `name`, `category`, `item`.

📄️ KnowledgeUpdate

Payload used to update Knowledge. Optional fields include `name`, `alias`, `type`, `collection_id`, `folder_id`, and 3 more.

📄️ LLMDetectedLimits

Attributes the endpoint reports for the requested ``model`` — token limits plus (Anthropic) reasoning-effort capability.

📄️ LLMEndpointProbeResult

Structured result of a single ``GET {base_url}/models`` probe.

📄️ LLMModelCreate

Payload used to create LLM Model. Required fields are `name`, `type`, `provider`.

📄️ LLMModelDeployment

Container deployment settings for an internally hosted LLM model.

📄️ LLMModelResponse

LLMModel API response. Never carries ``api_key``.

📄️ LLMModelTestRequest

External LLM endpoint connection test request. Supports both pre-save (form ``base_url`` + ``api_key``) and post-save (``model_id`` with ``api_key`` omitted → server loads the stored key). A body ``api_key`` always takes precedence over the stored one.

📄️ LLMNodeConfig

LLM workflow-node configuration that references a registered model. The registered model supplies the endpoint, credential, and generation defaults; this node supplies the prompt, output field, and optional reasoning effort.

📄️ LLMReasoningCapability

Whether a model exposes an OpenAI-style ``reasoning_effort`` control.

📄️ ListResponse[AccessControlPolicy]

Paginated response containing Access Control Policy records. The required field is `items`.

📄️ ListResponse[ActorResponse]

Paginated response containing Actor Response records. The required field is `items`.

📄️ ListResponse[AgentResponse]

Paginated response containing Agent Response records. The required field is `items`.

📄️ ListResponse[ArtifactTemplate]

Paginated response containing Artifact Template records. The required field is `items`.

📄️ ListResponse[ArtifactVersion]

Paginated response containing Artifact Version records. The required field is `items`.

📄️ ListResponse[AssistantLlmModelResponse]

Paginated response containing Assistant LLM Model Response records. The required field is `items`.

📄️ ListResponse[CodeResponse]

Paginated response containing Code Response records. The required field is `items`.

📄️ ListResponse[CollectionResponse]

Paginated response containing Collection Response records. The required field is `items`.

📄️ ListResponse[ConnectorSummaryResponse]

Paginated response containing Connector Summary Response records. The required field is `items`.

📄️ ListResponse[DashboardResponse]

Paginated response containing Dashboard Response records. The required field is `items`.

📄️ ListResponse[DatasetObjectEntry]

Paginated response containing Dataset Object Entry records. The required field is `items`.

📄️ ListResponse[DatasetResponse]

Paginated response containing Dataset Response records. The required field is `items`.

📄️ ListResponse[DocumentResponse]

Paginated response containing Document Response records. The required field is `items`.

📄️ ListResponse[EntityTypeResponse]

Paginated response containing Entity Type Response records. The required field is `items`.

📄️ ListResponse[FolderResponse]

Paginated response containing Folder Response records. The required field is `items`.

📄️ ListResponse[FunctionResponse]

Paginated response containing Function Response records. The required field is `items`.

📄️ ListResponse[GroupResponse]

Paginated response containing Group Response records. The required field is `items`.

📄️ ListResponse[KnowledgeResponse]

Paginated response containing Knowledge Response records. The required field is `items`.

📄️ ListResponse[LLMModelResponse]

Paginated response containing LLM Model Response records. The required field is `items`.

📄️ ListResponse[MarkingResponse]

Paginated response containing Marking Response records. The required field is `items`.

📄️ ListResponse[ModelResponse]

Paginated response containing Model Response records. The required field is `items`.

📄️ ListResponse[OIDCClientResponse]

Paginated response containing OIDC Client Response records. The required field is `items`.

📄️ ListResponse[PipelineResponse]

Paginated response containing Pipeline Response records. The required field is `items`.

📄️ ListResponse[PromptSummaryResponse]

Paginated response containing Prompt Summary Response records. The required field is `items`.

📄️ ListResponse[ReactionResponse]

Paginated response containing Reaction Response records. The required field is `items`.

📄️ ListResponse[RelationTypeResponse]

Paginated response containing Relation Type Response records. The required field is `items`.

📄️ ListResponse[ReportOutputResponse]

Paginated response containing Report Output Response records. The required field is `items`.

📄️ ListResponse[ReportResponse]

Paginated response containing Report Response records. The required field is `items`.

📄️ ListResponse[ReportRunResponse]

Paginated response containing Report Run Response records. The required field is `items`.

📄️ ListResponse[ReportTemplateResponse]

Paginated response containing Report Template Response records. The required field is `items`.

📄️ ListResponse[Runtime]

Paginated response containing Runtime records. The required field is `items`.

📄️ ListResponse[SecretSummaryResponse]

Paginated response containing Secret Summary Response records. The required field is `items`.

📄️ ListResponse[SemanticDictionaryResponse]

Paginated response containing Semantic Dictionary Response records. The required field is `items`.

📄️ ListResponse[SemanticTableResponse]

Paginated response containing Semantic Table Response records. The required field is `items`.

📄️ ListResponse[SemanticTemplateResponse]

Paginated response containing Semantic Template Response records. The required field is `items`.

📄️ ListResponse[SourceResponse]

Paginated response containing Source Response records. The required field is `items`.

📄️ ListResponse[ToolResponse]

Paginated response containing Tool Response records. The required field is `items`.

📄️ ListResponse[UserResponse]

Paginated response containing User Response records. The required field is `items`.

📄️ ListResponse[ViewResponse]

Paginated response containing View Response records. The required field is `items`.

📄️ ListResponse[str]

Paginated response containing str records. The required field is `items`.

📄️ LoginRequest

📄️ LoopNodeConfig

Settings that control Loop Node. The required field is `mode`.

📄️ MarkingCreate

Request to create a classification label in the global catalog. Administrator access is required, and classification labels are not scoped to a collection.

📄️ MarkingMemberGrant

Request body to grant marking clearance. Public is not a valid clearance holder — markings gate access explicitly, so a wildcard subject is rejected.

📄️ MarkingResourceRef

Resource reference accepted when applying a classification label. Users, groups, services, and classification labels themselves are not valid target resource types.

📄️ MarkingResponse

Marking catalog API response. A marking is a pure catalog entry, so it adds no fields beyond the common BaseResource set. ``owners``/``my_share_scope`` stay empty — a marking carries only a ``member`` clearance relation, not read/write/own roles.

📄️ MarkingUpdate

Marking update request. Markings stay global, so no ``collection_id``.

📄️ MemberAdd

Request model for adding a member to a group.

📄️ Message

API response message.

📄️ Metric

Reusable semantic metric. Simple and cumulative metrics require `agg`; ratio metrics require `numerator` and `denominator`; derived metrics require `expr` and `metrics`; conversion metrics require `base`, `conversion`, and `entity`.

📄️ MirrorHealthMarker

Degraded-mirror marker body.

📄️ ModelCreate

Payload used to create Model. Required fields are `name`, `type`, `protocol`.

📄️ ModelDeployment

Container deployment settings for an internally hosted model.

📄️ ModelExternalAuth

Authentication settings for an external model endpoint; secret values are referenced rather than returned.

📄️ ModelResponse

Data returned for Model. Required fields are `id`, `name`, `category`, `type`, `protocol`, `item`.

📄️ ModelToleration

Scheduling tolerance for an internally hosted model, allowing placement on a compute resource with a matching taint.

📄️ NotificationAction

Shared request and response schema. Request: Notification action accepted when configuring a reaction, including its recipients, message, and channel. Response: Notification action returned for a reaction, including its recipients, message, and channel.

📄️ NotificationEvent

Published notification event. `recipients_spec` preserves the original recipient predicates; the display `collection` may differ from `recipients_spec.collection`, which controls recipient routing.

📄️ NotificationPublish

Notification channel, message, payload, and recipients accepted when publishing a notification.

📄️ NotificationRecipients

Notification recipient predicates. Users, groups, collection readers, and broadcast recipients are combined and deduplicated. A request with no predicates is rejected with `422`; a valid predicate that resolves to no users creates zero deliveries.

📄️ OIDCClientCreate

OIDC client creation request. For confidential clients, the server generates a client_secret and returns the plaintext value once via OIDCClientSecretResponse.

📄️ OIDCClientResponse

OIDC client API response. Excludes client_secret_hash by design.

📄️ OIDCClientSecretResponse

Secret issuance response (returned only on create or rotate). The plaintext client_secret is shown exactly once. Callers must store it immediately; the server retains only the Argon2 hash.

📄️ OIDCClientUpdate

OIDC client update request. client_id, client_type, and grant_types are immutable after creation. Use the rotate-secret endpoint to change client_secret.

📄️ ObjectVersion

Stored-object version metadata, including its identifier, modification time, and latest-version status.

📄️ OntologyAction

Shared request and response schema. Request: Ontology write action accepted when configuring a reaction, including its target and field mappings. Response: Ontology write action returned for a reaction, including its target and field mappings.

📄️ OntologyBulkDeleteAccepted

Data returned by Trigger Ontology Bulk Delete. Required fields are `run_id`, `status`.

📄️ OntologyBulkDeleteCount

Aggregate progress by ontology resource kind.

📄️ OntologyBulkDeleteItem

A delete target captured at preview/operation start.

📄️ OntologyBulkDeleteItemStatus

Per-resource cleanup progress for a collection-scoped bulk delete.

📄️ OntologyBulkDeletePreview

Side-effect-free preview for collection ontology reset.

📄️ OntologyBulkDeleteRequest

`POST /ontology/collections/{id}/bulk-delete` body.

📄️ OntologyBulkDeleteStatus

Collection-scoped ontology bulk delete status.

📄️ OntologyCondition

Reaction condition that triggers when an entity or relation row change matches `target_id`, `op`, and the optional `filter`.

📄️ OntologyHealthResponse

Ontology consistency status for the complete set of ontology query surfaces, with resynchronization, orphan cleanup, and schema-drift details.

📄️ OntologyOrphanCleanupAccepted

Data returned by Trigger Ontology Orphan Cleanup. The required field is `run_id`.

📄️ OntologyOrphanCleanupRun

Active orphan-cleanup flow run summary.

📄️ OntologyResyncAccepted

Data returned by Trigger Ontology Resync. The required field is `run_id`.

📄️ OntologyResyncRequest

`POST /ontology/resync` body.

📄️ OntologyResyncRun

Active/last-known resync flow run summary.

📄️ OntologyTableResetDependency

Relation rows affected by an EntityType table reset.

📄️ OntologyTableResetPreview

Side-effect-free reset impact preview.

📄️ OntologyTableResetRequest

`POST.../table/reset` body for EntityType reset.

📄️ OntologyTableResetResponse

Result of clearing an ontology table, including the affected resource and deleted row counts. Related query and graph surfaces may reflect the reset asynchronously.

📄️ OrphanEntry

Orphan-table entry containing its schema and table names and first detection time.

📄️ OrphanMarker

Detected orphan-resource summary with affected resources and detection time.

📄️ OutputMapping

Workflow output mapping.

📄️ ParallelBranch

Parallel execution branch.

📄️ ParallelNodeConfig

Settings that control Parallel Node. The required field is `branches`.

📄️ ParentCondition

Reaction trigger that fires after a parent reaction succeeds or fails. Each hop increments `parent_depth`, and values greater than 16 are rejected to break cycles.

📄️ PasswordChange

Password change request.

📄️ PermissionGrant

Role granted to a user, group, service, or public subject for a resource.

📄️ PermissionGrantWithResource

Permission grant enriched with display metadata for read responses.

📄️ PersonalPromptResponse

``content``/``updated_at`` are null when the prompt is unset.

📄️ PersonalPromptUpdate

Body for ``PUT /auth/me/personal-prompt``.

📄️ PipelineCreate

Payload used to create Pipeline. Required fields are `name`, `type`, `collection_id`, `steps`.

📄️ PipelineData

Shared request and response schema. Request: Pipeline port reference accepted in a pipeline definition. Exactly one of `dataset`, `entity`, `relation`, `source`, `view`, or `variable` must be set. Sources and views are input-only and are rejected for output ports. Variables are limited to Python steps, cannot be named `event` or `objects`, and do not support `version`, `row_filters`, or `column_masks`. Response: Pipeline port reference returned in a pipeline definition. Exactly one of `dataset`, `entity`, `relation`, or `variable` must be set because `source` and `view` are input-only. Variables are limited to Python steps, cannot be named `event` or `objects`, and do not support `version`, `row_filters`, or `column_masks`.

📄️ PipelineResponse

Data returned for Pipeline. Required fields are `id`, `name`, `category`, `type`, `steps`, `item`.

📄️ PipelineStep

Shared request and response schema. Request: Pipeline step accepted in a definition. Exactly one of `script` or `transform` must be set. Transform configuration refers to `inputs` and `outputs` by port key, and transform steps cannot use `secrets`. Response: Pipeline step returned in a definition. Exactly one of `script` or `transform` is set. Transform configuration refers to `inputs` and `outputs` by port key, and transform steps do not use `secrets`.

📄️ PipelineUpdate

Payload used to update Pipeline. Optional fields include `name`, `alias`, `type`, `collection_id`, `folder_id`, and 5 more.

📄️ ProfileUpdate

Self-profile update request for /auth/me.

📄️ PromptCreate

Payload used to create Prompt. Required fields are `name`, `type`.

📄️ PromptLongContext

A 1:1 uploaded document converted to Markdown for agent prompts.

📄️ PromptRenderRequest

Render request: variable name -> bound value.

📄️ PromptRenderResponse

Rendered prompt content (preview).

📄️ PromptResponse

Data returned for Prompt. Required fields are `id`, `name`, `category`, `type`, `item`.

📄️ PromptSummaryResponse

Prompt list/selector response (``content`` omitted for payload size).

📄️ PromptUpdate

Prompt update request (all fields optional; omitted = no change).

📄️ PromptUploadResponse

Prompt upload accepted for asynchronous Markdown conversion.

📄️ PromptVariable

A declared ``{{name}}`` template variable. Declaration only — binding values are supplied by the referencing agent; the Prompt itself stores no values.

📄️ ProviderInfo

Provider information.

📄️ PythonFunctionSchema

Python function schema definition.

📄️ QueryRequest

Table-query request containing SQL text and an optional row limit.

📄️ QueryResult

Cypher query execution result.

📄️ QuickUploadResponse

Data returned for Quick Upload. Required fields are `results`, `message`.

📄️ QuickUploadResult

Individual file upload result.

📄️ ReactionCreate

Payload used to create Reaction. Required fields are `name`, `collection_id`, `condition`, `actions`.

📄️ ReactionResponse

Reaction resource returned by the API, including resource metadata, its trigger condition, and its ordered action chain.

📄️ ReactionUpdate

Partial reaction update. Omitted fields keep their current values.

📄️ ReferencesResponse

Data returned for References. Required fields are `references`, `has_references`.

📄️ RelationTypeCreate

Payload used to create Relation Type. Required fields are `name`, `collection_id`, `source`, `target`.

📄️ RelationTypeResponse

Data returned for Relation Type. Required fields are `id`, `name`, `category`, `source`, `target`, `item`.

📄️ RelationTypeUpdate

Partial update for a relation type. `name` must match the current name, and schema changes must be additive.

📄️ RenameCastColumn

One column transform in ``rename_cast``: an optional rename (``alias``) and/or an allowlisted ``cast``. At least one must be set.

📄️ RenameCastConfig

Config for the ``rename_cast`` op. Listed columns are renamed / cast; columns not listed pass through unchanged. ``input`` is a port key.

📄️ RenameCastTransform

The ``rename_cast`` member of the ``PipelineTransform`` union.

📄️ ReportArtifactRef

Generated or planned output artifact reference.

📄️ ReportCreate

Payload used to create Report. The required field is `name`.

📄️ ReportInputRef

Object-storage input reference for report generation.

📄️ ReportOutputCreate

Payload used to create Report Output. Required fields are `report_id`, `run_id`, `format`, `role`, `mime`.

📄️ ReportOutputResponse

Data returned for Report Output. Required fields are `id`, `name`, `category`, `report_id`, `run_id`, `format`, and 3 more.

📄️ ReportOutputUpdate

Payload used to update Report Output. Optional fields include `name`, `alias`, `collection_id`, `folder_id`, `tags`, and 8 more.

📄️ ReportPlanRef

Intermediate plan artifact reference for a report run.

📄️ ReportResponse

Data returned for Report. Required fields are `id`, `name`, `category`, `item`.

📄️ ReportRunCreate

Payload used to create Report Run. The required field is `prompt`.

📄️ ReportRunResponse

Data returned for Report Run. Required fields are `id`, `name`, `category`, `report_id`, `status`, `prompt`, and 1 more.

📄️ ReportRunStatusTransition

Conditional report run status transition request.

📄️ ReportRunUpdate

Payload used to update Report Run. Optional fields include `name`, `alias`, `tags`, `description`.

📄️ ReportSourceFinalizationRequest

Atomically freeze canonical sources and their staged snapshots.

📄️ ReportSourceProvenance

Audit metadata for the immutable bytes used by a report run.

📄️ ReportTemplateCreate

Payload used to create Report Template. The required field is `name`.

📄️ ReportTemplateResponse

Data returned for Report Template. Required fields are `id`, `name`, `category`, `item`.

📄️ ReportTemplateUpdate

Payload used to update Report Template. Optional fields include `name`, `alias`, `collection_id`, `folder_id`, `tags`, and 7 more.

📄️ ReportUpdate

Payload used to update Report. Optional fields include `name`, `alias`, `collection_id`, `folder_id`, `tags`, and 5 more.

📄️ ResourceReference

Cross-collection resource reference.

📄️ ResourceVersion

Resource history record with version identity, latest-version status, timestamps, operation, author, change summary, and the underlying storage reference.

📄️ Role

Allowed Role values. Values: `owner`, `writer`, `reader`, `use`.

📄️ RoutedTable

One routed semantic table. ``score`` is the bi-encoder cosine; ``rerank_score`` (0-1) is the cross-encoder score, present only on reranked (``routed``) entries.

📄️ RoutedTemplate

Semantic template selected for a request with its match and ranking scores.

📄️ RowFilter

Shared request and response schema. Request: Nested row-filter group accepted in a request. Conditions are combined with `and` or `or`. Response: Nested row-filter group returned by the API. Conditions are combined with `and` or `or`.

📄️ RowFilterCondition

Leaf node of a row-filter expression (kind='condition').

📄️ RowFilterOperator

Allowed Row Filter Operator values. Values: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`, `in`, `not_in`, and 4 more.

📄️ Runtime

Runtime image available to internally hosted models, including its name, image reference, base environment, and notable included packages.

📄️ SecretBinding

Runtime secret injection request. `ref` identifies a stored secret and `name` selects its key under `contexts["secrets"]`; the execution principal must have `use` permission.

📄️ SecretCreate

Payload used to create Secret. The required field is `name`.

📄️ SecretResponse

Secret API response. Values are write-only and never returned.

📄️ SecretSummaryResponse

Data returned for Secret Summary. Required fields are `id`, `name`, `category`, `item`.

📄️ SecretUpdate

Secret update request. ``values`` is a partial merge/upsert.

📄️ SecretValueInput

Request-only secret value (write-only; never returned).

📄️ SelectColumn

Column selected by a `select_filter` transform. `alias` may rename the output column; this transform does not cast types, so use `rename_cast` when a type change is required.

📄️ SelectFilterConfig

Shared request and response schema. Request: Select-and-filter transform configuration accepted in a request. `input` is a pipeline port key, not a resource identifier; `columns=null` keeps every column, and `where=null` applies no predicate. Response: Select-and-filter transform configuration returned by the API. `input` is a pipeline port key, not a resource identifier; `columns=null` keeps every column, and `where=null` applies no predicate.

📄️ SelectFilterTransform

Shared request and response schema. Request: Select-and-filter transform accepted as a pipeline step. Response: Select-and-filter transform returned as part of a pipeline step.

📄️ SemanticDictionaryCreate

Payload used to create Semantic Dictionary. The required field is `name`.

📄️ SemanticDictionaryResponse

Semantic dictionary resource returned with its configuration and item count.

📄️ SemanticDictionaryUpdate

SemanticDictionary update request (all fields optional; omitted = no change).

📄️ SemanticRoutingSearchRequest

Preview request for table/template routing — POST /semantic/agents/{agent_id}/{tables,templates}/search.

📄️ SemanticTableCreate

Payload used to create Semantic Table. The required field is `name`.

📄️ SemanticTableResponse

Data returned for Semantic Table. Required fields are `id`, `name`, `category`, `item`.

📄️ SemanticTableUpdate

SemanticTable update request (all fields optional; omitted = no change).

📄️ SemanticTemplateCreate

Payload used to create Semantic Template. The required field is `name`.

📄️ SemanticTemplateResponse

SemanticTemplate detail/list response (verified query template group).

📄️ SemanticTemplateUpdate

SemanticTemplate update request (all fields optional; omitted = no change). ``templates`` replaces the whole item list.

📄️ ServerInfo

Graph service connection address, protocol version, and server agent information.

📄️ ServiceTokenCreateRequest

Request to create a service token for a service account.

📄️ ServiceTokenCreateResponse

Response after creating a service token. Token is shown only once.

📄️ ServiceTokenListItem

Service token metadata for listing (excludes sensitive fields).

📄️ SourceColumnsResponse

Column names exposed by a source; the list is empty when the source cannot be read.

📄️ SourceCreate

Payload used to create Source. Required fields are `name`, `type`, `collection_id`.

📄️ SourceResponse

Data returned for Source. Required fields are `id`, `name`, `category`, `type`, `item`.

📄️ SourceUpdate

Partial update for a source. Do not include `name`; `collection_id` may move the source to another collection but cannot be cleared.

📄️ SqlCheckRequest

SQL access-check request containing the query and optional SQL dialect.

📄️ SqlCheckResponse

SQL access-check result indicating whether the query is allowed and returning the original or policy-rewritten SQL when allowed.

📄️ StartNodeConfig

Settings that control Start Node. Optional fields include `variables`.

📄️ SubjectLookupRequest

`POST /directory/subjects:lookup` request body.

📄️ SubjectLookupResponse

`POST /directory/subjects:lookup` response envelope.

📄️ SubjectSearchResponse

Ranked directory-search results visible to the caller. `next_cursor` is always null in the current response contract.

📄️ SubjectSummary

Subject identity returned by the API, including its identifier, type, display name, and optional avatar data. `email` is returned only when `type` is `user`.

📄️ SubjectType

Allowed Subject Type values. Values: `user`, `group`, `service`, `public`.

📄️ TableRoutingResponse

Semantic-table routing result containing the request, candidate tables, and selected tables.

📄️ TableVersion

Table version metadata including its version number, timestamp, latest-version status, operation, author, and change summary when available.

📄️ TemplateItem

Verified reusable computation template with example questions, shared slots, named queries, and code that combines query results. A single-query template uses one query with pass-through code.

📄️ TemplateNodeConfig

Settings that control Template Node. The required field is `template`.

📄️ TemplateQuery

One sub-query within a compute TemplateItem. ``name`` is the key its rows land under in ``run(frames)`` (unique within the item); ``table_id`` binds it to the SemanticTable it compiles against; ``query`` is the parameterized query-object JSON (``QueryMetrics``/``QueryRows`` shape per ``tool``) with ``{{slot}}`` placeholders.

📄️ TemplateRoutingResponse

Semantic-template routing result containing the request, candidate templates, and selected templates.

📄️ TemplateSlot

A declared template variable. ``kind`` guides the filler: ``value`` marks a natural-language surface form resolved by value linking at execution time (write it as the user says it); ``number`` substitutes an int/float; ``list`` takes an array and fans each sub-query that references it (via an ``op="="`` condition) into one sub-query per element, named ``{sub-query}__{element}`` — the item's ``code`` aggregates across them.

📄️ TimeCondition

Time-based reaction trigger whose `schedule` is a standard five-field cron expression.

📄️ TokenResponse

Data returned for Token. The required field is `access_token`.

📄️ ToolCallNodeConfig

Settings that control Tool Call Node. The required field is `tool`.

📄️ ToolCreate

Payload used to create Tool. Required fields are `name`, `description`, `type`.

📄️ ToolDuplicate

Tool duplicate (save-as / clone) request. The source tool's executable definition (type/function/ref/options) is copied verbatim by the server; only identity/placement metadata is overridable here. The server always mints a fresh id and forces is_builtin=False, so a clone of a read-only builtin lands as an editable user tool.