DocumentationAI DocumentationAI Assistant in Adobe Experience Cloud

Adobe Experience Platform tools in Adobe CX Enterprise MCP aep-mcp

Last update: June 10, 2026

CREATED FOR:

  • Admin
  • User
  • Developer
  • Leader

You can use the Adobe Experience Platform product tools to inspect schemas, datasets, data governance configuration, Query Service resources, and audit events from an MCP-compatible client. These tools are available through the Adobe CX Enterprise MCP when your organization is enabled and your user account has the required Experience Platform permissions.

AVAILABILITY
The Experience Platform product tool is in Beta. Access is by invitation only and requires Adobe organization enablement. See Access CX Enterprise MCP tools.

Summary

Tool
Description
Resource
Capabilities
Status
search_allowed_ip_ranges
Retrieve Query Service IP access restrictions
Data Distiller Auth · IP ranges
list
Active
search_audit
List user activity audit events across Experience Platform
Audit Query · audit events
list, filter by asset type, action, status, time range
Active
search_datasets
Query dataset and batch ingestion metadata
Catalog API · dataSets, batches
list, get, filter, list last, list files
Active
search_class_relations
Search Experience Platform business class relationships
Class Relations · static YAML index
search by token, multi-term, partial match
Active
search_data_access
List files from failed ingestion batches
Data Access API · failed batches
list failed files
Active
search_data_lake
Inspect dataset metadata and batch health
Data Lake API · datasets, batches
get, get size, list failed batches
Active
search_dule
Query data governance labels, policies, actions
Data Governance · labels, policies, marketing_actions
list, get, list enabled, evaluate
Active
search_query_service
Query SQL queries, templates, schedules, alerts
Query Service · queries, templates, schedules, alerts
list, get, filter, get connection params
Active
search_schema_registry
Query XDM schemas, field groups, classes, types
Schema Registry · schemas, fieldgroups, classes, data_types, descriptors
list, get, filter by container
Active

Tool Reference

search_allowed_ip_ranges

Resource: Data Distiller Auth · IP ranges
Status: Active

Retrieve all configured IP access restrictions for Query Service in the current sandbox. Returns the organization ID and the list of allowed IP ranges. Only available to customers with the Data Distiller add-on.

Capabilities: list allowed IP ranges for Query Service

No parameters.

search_audit

Resource: Audit Query · audit events
Status: Active

List timestamped records of user activities across Experience Platform services. Returns action type, user email, asset information, and event status. Use asset_type and action to narrow results. Defaults to the last 7 days when no time range is specified. Limited to the last 1000 records and events from the past 90 days.

Capabilities: list audit events, filter by asset type, action, status, time range, paginate

Parameters:

Parameter
Required
Description
action
No
Filter by action type. Common values (comma-separate for OR): Create, Delete, Update, Enable, Disable
asset_type
No
Filter by asset type. Must be one of: Dataset, Schema, Segment, Destination, Source Data Flow, Merge Policy, Identity Namespace, Identity Graph, Sandbox, Role, Query, Scheduled Query, Datastream, Computed Attribute, Field Group, Class, Data Types, Account, Product Profile, Query Template, Work Order, Audit Logs, Access Control Policy
status
No
Filter by event status. Values: Success, Failure, Allow, Deny. Comma-separate for OR
start_time
No
Earliest timestamp. ISO 8601 UTC with ms, e.g. 2024-01-15T00:00:00.000Z
end_time
No
Latest timestamp. ISO 8601 UTC with ms
property_filter
No
Raw filter expression, e.g. action==create. Prefer the dedicated parameters above
orderby
No
Sort order: timestamp (asc) or -timestamp (desc)
limit
No
Maximum number of results (3–1000, default 50)
start
No
Pagination offset. Increment by limit value for each page
query_id
No
Query ID from a previous response to repeat the same query

search_datasets

Resource: Catalog API · dataSets, batches
Status: Active

Unified dispatch tool for the Experience Platform Catalog Service. Query dataset metadata (schema refs, tags, creation info) or batch ingestion records (status, metrics, file listings). Use dataset/list to discover datasets, batch/list to check ingestion health, and batch/list_files or batch/get_meta_files to inspect specific batch contents. All operations are read-only.

Capabilities: list datasets, get dataset, list batches, get batch, list last batch per dataset, list batch files, get batch meta files (row errors, input files)

Parameters:

Parameter
Required
Description
entity_type
Yes
dataset or batch
operation
Yes
list, get, list_last, list_files, get_meta_files. Valid combos: dataset → list, get; batch → all five
resource_id
No
Dataset or batch ID. Required for dataset/get, batch/get, batch/list_files, batch/get_meta_files
query_params.limit
No
Max results per page (max 100). Applies to all list operations
query_params.start
No
Pagination offset. Applies to all list operations
query_params.order_by
No
Sort direction, e.g. asc:created,updated. Applies to all list operations
query_params.properties
No
Comma-separated property allowlist. Applies to dataset/list, dataset/get, batch/list, batch/list_last
query_params.name
No
Filter datasets by name (dataset/list only)
query_params.tags
No
Filter datasets by tags, e.g. unifiedProfile:enabled:true (dataset/list only)
query_params.property_filter
No
Regex filter on response objects (dataset/list and batch/list)
query_params.status
No
Filter batches by status: success, failed, loading, active (batch/list only)
query_params.dataset_id
No
Scope batches to a specific dataset (batch/list and batch/list_last)
query_params.created_after
No
Filter batches created after Unix timestamp in ms (batch/list only)
query_params.created_before
No
Filter batches created before Unix timestamp in ms (batch/list only)
query_params.last_batch_status
No
Filter by last batch status (batch/list_last only)
query_params.aggregate
No
Return aggregated metrics at root level (batch/get only)
query_params.path
No
Meta file to download: row_errors, input_files, row_errors_sample.json (batch/get_meta_files only)

search_class_relations

Resource: Class Relations · static YAML index
Status: Active

Search for Experience Platofrm business class relationships by name using the static class_relations_v1.yaml index. No Experience Platofrm API calls are made. Accepts a single term or comma-separated terms; each term is matched against class names using partial token matching. Returns matching classes with forward relations (what each class points to) and reverse relations (which classes point back to it). Use this to understand entity relationships before building queries, dataflows, or schema compositions.

Capabilities: search by token, multi-term comma-separated search, partial token match, bidirectional synonym expansion

Parameters:

Parameter
Required
Description
query
Yes
Business class name or object type to search for. Supports partial token matches (dat matches dataset, data_type, etc.). Pass multiple comma-separated terms to search for several classes at once (e.g. dataset, schema)
n
No
Maximum number of matched results to return (default 5, min 1)

search_data_access

Resource: Data Access API · failed batches
Status: Active

Access files from failed Experience Platofrm data ingestion batches. Use failed_batch/list_failed to list the files belonging to a failed batch for failure diagnosis. Requires a batch ID for all operations. Note: file/get and dataset/preview are disabled as they expose actual record data. All operations are read-only.

Capabilities: list files from a failed ingestion batch

Parameters:

Parameter
Required
Description
entity_type
Yes
failed_batch — list files from a failed ingestion batch
operation
Yes
list_failed — the only supported operation
resource_id
Yes
Batch ID of the failed batch
query_params.start
No
Paging start index, e.g. 1
query_params.limit
No
Number of results per page, e.g. 10
query_params.path
No
Full file name filter, e.g. profiles.csv

search_data_lake

Resource: Data Lake API · datasets, batches
Status: Active

Inspect dataset and batch metadata from the Data Lake layer. Use get for full metadata, get_size for storage and ingestion size metrics, and list_failed to monitor ingestion failures within a time window. Defaults to the last 7 days when no time range is provided for list_failed. All operations are read-only and require a resource ID.

Capabilities: get dataset/batch metadata, get storage size metrics, list failed batches within a time window

Parameters:

Parameter
Required
Description
entity_type
Yes
dataset or batch
operation
Yes
get, get_size, list_failed. list_failed only supports batch entity type
resource_id
Yes
Dataset ID or batch ID. For list_failed: the dataset ID to scope failures to
query_params.created_after
No
Start of time window. Unix timestamp in ms
query_params.created_before
No
End of time window. Unix timestamp in ms
query_params.limit
No
Max results per page (max 100)
query_params.order_by
No
Sort direction, e.g. desc:created

search_dule

Resource: Data Governance · labels, policies, marketing_actions
Status: Active

Query the Policy Service API for data usage labels, policies, and marketing actions. Use marketing_action/evaluate to test whether a marketing action on data with specific labels would violate any governance policies. All operations are read-only.

Capabilities: list/get data usage labels, list/get policies, list enabled policies, list/get marketing actions, evaluate marketing action against labels

Parameters:

Parameter
Required
Description
entity_type
Yes
label, policy, or marketing_action
operation
Yes
list, get, list_enabled (policy only), evaluate (marketing_action only). list_enabled does not require scope
scope
No
core (Adobe-defined) or custom (org-defined). Required for list, get, evaluate; not used for list_enabled
resource_id
No
Label name, policy ID, or marketing action name. Required for get and evaluate
query_params.dule_labels
No
Comma-separated labels (e.g. C1,C3). Required for marketing_action/evaluate; optional filter for policy/list
query_params.limit
No
Max results
query_params.start
No
Pagination cursor from a previous response’s _page.next value
query_params.orderby
No
Comma-separated sort fields
query_params.property_filter
No
Filter expression, e.g. name==C1
query_params.marketing_action
No
Restrict policy list to policies referencing this marketing action (policy/list only)
query_params.include_draft
No
Include DRAFT policies in marketing_action/evaluate (default: ENABLED policies only)

search_query_service

Resource: Query Service · queries, templates, schedules, schedule runs, connections, alert subscriptions
Status: Active

Unified tool for Query Service resources. List and retrieve ad-hoc queries, saved SQL templates, scheduled queries and their runs, interactive connection parameters (for psql/JDBC clients), and alert subscriptions. For query lists, default to isService==false,isParentLevel==true to filter out internal traffic. All operations are read-only.

Capabilities: list/get queries, list/get templates, list/get schedules, list/get schedule runs, get connection params, list alert subscriptions

Parameters:

Parameter
Required
Description
entity_type
Yes
query, query_template, schedule, schedule_run, connection, alert_subscription
operation
Yes
list, get, get_connection_params, list_by_u...
recommendation-more-help
experience-cloud-ai-help