Azure MCP Server tools for Azure Data Manager for Energy

By using Azure MCP Server, you can retrieve or search for OSDU records or schemas in Azure Data Manager for Energy by using natural language prompts.

Azure Data Manager for Energy is a managed cloud data platform for the energy industry. It combines the capabilities of the OSDU® Data Platform with Azure services to help organizations manage, store, search, and share energy data.

Note

Tool parameters: The Azure MCP Server tools define parameters for data they need to complete tasks. Some of these parameters are specific to each tool and are documented here. Other parameters are global and shared by all tools. For more information, see Tool parameters.

Health: Check

Checks an ADME/OSDU endpoint for health, authentication, and connectivity. Returns the service health status, detailed error information, and the HTTP status code.

Azure MCP CLI

Example CLI command

azmcp adme health check --endpoint <endpoint> \
                         --data-partition <data-partition>

Parameter	Type	Required	Description
endpoint	string	Yes	The service endpoint, for example https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, for example contoso-dp1.

MCP Server

Example prompts include:

• "Check the health, authentication, and connectivity of the Azure Data Manager for Energy endpoint https://contoso.energy.azure.com for data partition contoso-dp1."

Parameter	Required or optional	Description
Endpoint	Required	The service endpoint, for example https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, for example contoso-dp1.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Schema: Get

Gets one ADME/OSDU schema by exact Kind, and returns the full JSON definition, including fields, property types, and structure. It's useful to inspect nested types, arrays, and property metadata to validate schema design or map incoming data.

Azure MCP CLI

Example CLI command

azmcp adme schema get --endpoint <endpoint> \
                       --data-partition <data-partition> \
                       --kind <authority:source:entity-type:version> 

Parameter	Type	Required	Description
kind	string	Yes	The fully-qualified kind authority:source:type:version, for example osdu:wks:master-data--Well:1.0.0. Wildcards aren't supported. Use azmcp adme schema list to discover valid kinds and versions.
endpoint	string	Yes	The service endpoint, for example https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, for example contoso-dp1.

MCP Server

Example prompts include:

• "Get the schema osdu:wks:master-data--Well:1.0.0 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Show the fields, property types, and structure of the OSDU schema osdu:wks:work-product-component--WellLog:1.0.0 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Retrieve the full JSON definition for the OSDU schema osdu:wks:master-data--Wellbore:1.0.0 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
Kind	Required	The fully qualified kind authority:source:type:version, for example osdu:wks:master-data--Well:1.0.0. Wildcards aren't supported. Use azmcp adme schema list to discover valid kinds and versions.
Endpoint	Required	The service endpoint, for example https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, for example contoso-dp1.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Schema: List

List multiple ADME and OSDU schema descriptors. Use optional filters for authority, source, entity type, status, scope, schema version, latest version, offset, and limit. The command returns descriptors that match the specified filters.

Azure MCP CLI

Example CLI command

azmcp adme schema list --endpoint <endpoint> \
                        --data-partition <data-partition> \
                        [--authority <authority>] \
                        [--source <source>] \
                        [--entity-type <entity-type>] \
                        [--status <PUBLISHED|DEVELOPMENT|OBSOLETE>] \
                        [--scope <SHARED|INTERNAL>] \
                        [--schema-version-major <major>] \
                        [--schema-version-minor <minor>] \
                        [--schema-version-patch <patch>] \
                        [--latest-version <TRUE|FALSE>] \
                        [--offset <offset>] \
                        [--limit <limit>]

Parameter	Type	Required	Description
endpoint	string	Yes	The service endpoint, for example https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, for example contoso-dp1.
authority	string	No	Filter by the authority segment of the kind, for example osdu in osdu:wks:master-data--Well:1.0.0.
source	string	No	Filter by the source segment of the kind, for example wks in osdu:wks:master-data--Well:1.0.0.
entity-type	string	No	Filter by the entity type segment of the kind, for example master-data--Well or work-product-component--WellLog.
status	enumeration	No	Filter by lifecycle status: PUBLISHED, DEVELOPMENT, or OBSOLETE. Omit to return schemas in all lifecycle statuses.
scope	enumeration	No	Filter by scope: SHARED for system-defined schemas, INTERNAL for schemas defined in this data partition. Omit to return both.
schema-version-major	integer	No	Filter by schema major version, for example 1. When combined with --latest-version, supply version filters in order (major, then minor).
schema-version-minor	integer	No	Filter by schema minor version, for example 0. When --latest-version is true, requires --schema-version-major.
schema-version-patch	integer	No	Filter by schema patch version, for example 0. When --latest-version is true, requires --schema-version-major and --schema-version-minor.
latest-version	boolean	No	Return only the newest version of each schema entity, collapsing duplicates across scopes. If the switch is included without a value, it defaults to true. When filtering by version, supply components in order: major, then minor, then patch.
offset	integer	No	The starting offset for paging; compare with the response's totalCount to decide whether to fetch further pages.
limit	integer	No	The number of schema descriptors to return in one page.

MCP Server

Example prompts include:

• "List all shared Well schemas from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "List the latest published shared OSDU schemas for entity type master-data--Well from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "List development OSDU schemas from authority osdu and source wks at endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
Endpoint	Required	The service endpoint, for example https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, for example contoso-dp1.
Authority	Optional	Filter by the authority segment of the kind, for example osdu in osdu:wks:master-data--Well:1.0.0.
Source	Optional	Filter by the source segment of the kind, for example wks in osdu:wks:master-data--Well:1.0.0.
Entity type	Optional	Filter by the entity type segment of the kind, for example master-data--Well or work-product-component--WellLog.
Status	Optional	Filter by lifecycle status: PUBLISHED, DEVELOPMENT, or OBSOLETE. Omit to return schemas in all lifecycle statuses.
Scope	Optional	Filter by scope: SHARED for system-defined schemas, INTERNAL for schemas defined in this data partition. Omit to return both.
Schema version major	Optional	Filter by schema major version, for example 1. When combined with --latest-version, supply version filters in order (major, then minor).
Schema version minor	Optional	Filter by schema minor version, for example 0. When --latest-version is true, requires --schema-version-major.
Schema version patch	Optional	Filter by schema patch version, for example 0. When --latest-version is true, requires --schema-version-major and --schema-version-minor.
Latest version	Optional	Return only the newest version of each schema entity, collapsing duplicates across scopes. When filtering by version, supply components in order: major, then minor, then patch.
Offset	Optional	The starting offset for paging. Compare with the response's totalCount to decide whether to fetch further pages.
Limit	Optional	The number of schema descriptors to return in one page.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Search

Searches ADME/OSDU records across one or more exact or wildcard kinds, using indexed criteria. Returns full records, selected fields, or aggregate counts.

Azure MCP CLI

Example CLI command

azmcp adme search --endpoint <endpoint> \
                  --data-partition <data-partition> \
                  --kind <authority:source:entity-type:version> [<kind>...] \
                  [--query <lucene-query>] \
                  [--limit <limit>] \
                  [--cursor-pagination-mode <TRUE|FALSE>] \
                  [--search-after <TRUE|FALSE>] \
                  [--cursor <cursor>] \
                  [--offset <offset>] \
                  [--returned-fields <path> [<path>...]] \
                  [--aggregate-by <path>] \
                  [--track-total-count <TRUE|FALSE>] \
                  [--sort <json-object>] \
                  [--spatial-filter <json-object>] \
                  [--query-as-owner <TRUE|FALSE>] \
                  [--excluded-fields <path> [<path>...]] \
                  [--highlighted-fields <path> [<path>...]] \
                  [--suggest-phrase <phrase>]

Parameter	Type	Required	Description
kind	array of strings	Yes	One or more kind selectors in authority:source:type:version format with optional * wildcards, for example ["osdu:wks:master-data--Well:1.0.0"]. Several kinds are searched as a union. Use the narrowest kind; each * widens the scan and its cost.
endpoint	string	Yes	The service endpoint, for example https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, for example contoso-dp1.
query	string	No	Optional Lucene filter over indexed fields, for example data.FieldID:Volve*. Prefix payload fields with data. and add .keyword for exact, case-sensitive, or null matches. Quote full record IDs. Trailing wildcards are supported; leading wildcard forms such as *Volve and *Volve* are rejected by ADME. Omit to match all records of the kind.
limit	integer	No	Page size (1-1000). Defaults to 10. Pair a large page with returnedFields.
cursor-pagination-mode	boolean	No	Use cursor pagination for a point-in-time snapshot, bulk processing, or more than 10000 results. If the switch is omitted, it defaults to false. If the switch is included without a value, it defaults to true. A supplied cursor also selects cursor pagination.
search-after	boolean	No	Allow index changes to affect later cursor pages instead of using a point-in-time snapshot. If the switch is included without a value, it defaults to true. Keep enabled on continuation requests.
offset	integer	No	Starting offset for query pagination. Cannot be combined with cursor pagination, and offset + limit must be <= 10000.
cursor	string	No	Continuation token from a previous cursor response. Supplying it selects cursor pagination; continue until results is empty. Resend all original cursor criteria unchanged because the service does not reliably retain page settings. If a continuation fails, restart from the first page to avoid missing records. Cannot be combined with offset or aggregateBy and expires after about 1 minute.
returned-fields	array of strings	No	Optional field paths to project, for example ["id","kind","data.FacilityName"]. Omit to return full records.
aggregate-by	string	No	Optional field to aggregate on for distinct values/counts, for example kind or data.FacilityName.keyword. Uses query pagination and cannot be combined with cursor.
track-total-count	boolean	No	Set true to request an exact totalCount. If the switch is included without a value, it defaults to true. Cursor searches always return an exact total count.
sort	string	No	Optional sort criteria. Provide a JSON object with equal-length field and order arrays, for example {"field":["data.Name.keyword","id"],"order":["DESC","ASC"]}.
spatial-filter	string	No	Geo-spatial filter (for example, byBoundingBox or byDistance). Pass a JSON object. A field that is not a geo-point yields 0 results rather than an error.
query-as-owner	boolean	No	If true, returns only records the user owns. If the switch is omitted, it defaults to false. If the switch is included without a value, it defaults to true.
excluded-fields	array of strings	No	Optional fields to exclude from the payload; may be ignored by the service, so prefer returnedFields.
highlighted-fields	array of strings	No	Optional fields to highlight with matched snippets in the response.
suggest-phrase	string	No	Optional phrase for did you mean spell-check; availability depends on service configuration.

MCP Server

Example prompts include:

• "Search records of kind osdu:wks:master-data--Well:1.0.0 that match the indexed-field Lucene filter data.FacilityName:Volve* at endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Search the kinds osdu:wks:master-data--Well:* and osdu:wks:master-data--Wellbore:* for records that match data.Name:Volve*, and return only id, kind, and data.Name from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Search more than 10,000 records of kind osdu:wks:master-data--Wellbore:* as a cursor-paginated point-in-time snapshot, and return only id, kind, and data.Name from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Continue the snapshot search by using the cursor from the previous response and resending the original kind osdu:wks:master-data--Wellbore:*, limit, and returned fields id, kind, and data.Name unchanged for endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Search records of kind osdu:wks:master-data--Well:1.0.0 inside the specified bounding box and sort the results by id in descending order at endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
Kind	Required	One or more kind selectors in authority:source:type:version format with optional * wildcards, for example ["osdu:wks:master-data--Well:1.0.0"]. Several kinds are searched as a union. Use the narrowest kind because each * widens the scan and its cost.
Endpoint	Required	The service endpoint, for example https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, for example contoso-dp1.
Query	Optional	Optional Lucene filter over indexed fields, for example data.FieldID:Volve*. Prefix payload fields with data. and add .keyword for exact, case-sensitive, or null matches. Quote full record IDs. Trailing wildcards are supported, but ADME rejects leading wildcard forms such as *Volve and *Volve*. Omit to match all records of the kind.
Limit	Optional	Page size (1-1000). Defaults to 10. Pair a large page with returnedFields.
Cursor pagination mode	Optional	Use cursor pagination for a point-in-time snapshot, bulk processing, or more than 10000 results. Defaults to false for real-time query pagination. A supplied cursor also selects cursor pagination.
Search after	Optional	Allow index changes to affect later cursor pages instead of using a point-in-time snapshot. Keep enabled on continuation requests.
Offset	Optional	Starting offset for query pagination. It can't be combined with cursor pagination, and offset + limit must be less than or equal to 10,000.
Cursor	Optional	Continuation token from a previous cursor response. Supplying it selects cursor pagination. Continue until results is empty. Resend all original cursor criteria unchanged because the service doesn't reliably retain page settings. If a continuation fails, restart from the first page to avoid missing records. The cursor can't be combined with offset or aggregateBy and expires after about 1 minute.
Returned fields	Optional	Optional field paths to project, for example ["id","kind","data.FacilityName"]. Omit to return full records.
Aggregate by	Optional	Optional field to aggregate on for distinct values or counts, for example kind or data.FacilityName.keyword. Uses query pagination and can't be combined with cursor.
Track total count	Optional	Set true to request an exact totalCount. Cursor searches always return an exact total count.
Sort	Optional	Optional sort criteria. Provide a JSON object with equal-length field and order arrays, for example {"field":["data.Name.keyword","id"],"order":["DESC","ASC"]}.
Spatial filter	Optional	Geospatial filter, such as byBoundingBox or byDistance. Pass a JSON object. A field that isn't a geographic point returns no results instead of an error.
Query as owner	Optional	If true, returns only records the user owns. Defaults to false.
Excluded fields	Optional	Optional fields to exclude from the payload; may be ignored by the service, so prefer returnedFields.
Highlighted fields	Optional	Optional fields to highlight with matched snippets in the response.
Suggest phrase	Optional	Optional phrase for a "did you mean" spelling check. Availability depends on the service configuration.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	❌	❌	✅	❌	❌

Storage record: Fetch

Fetches multiple ADME/OSDU records in a single batch by using record IDs. Returns full record content, selected attributes, or conversion status and errors.

Azure MCP CLI

Example CLI command

azmcp adme storage record fetch --endpoint <endpoint> \
                                  --data-partition <data-partition> \
                                  --ids <id> [<id>...] \
                                  [--attributes <path> [<path>...]] \
                                  [--frame-of-reference <TRUE|FALSE>]

Parameter	Type	Required	Description
ids	array of strings	Yes	The fully qualified record IDs to fetch, each in the format {partition}:{object-type}:{unique-id}. Specify up to 20 IDs without --attributes, or up to 100 with it.
endpoint	string	Yes	The service endpoint, such as https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, such as contoso-dp1.
attributes	array of strings	No	Dotted-path fields to return instead of whole records, for example data.Name and data.FileSourceInfo. Use it only when specific fields are requested; it can't be combined with --frame-of-reference.
frame-of-reference	boolean	No	Convert measurements to SI, coordinates to WGS84, and dates to UTC on the server. If the switch is included without a value, it defaults to true. The response then includes conversionStatuses, which is empty for records without measured or spatial fields.

MCP Server

Example prompts include:

• "Fetch records opendes:well:W-99 and opendes:well:W-100 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Fetch the full content of records opendes:well:W-99, opendes:well:W-100, and opendes:well:W-101 in one batch from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Fetch only data.Name and data.Description for records opendes:well:W-99 and opendes:well:W-100 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Fetch record opendes:well:W-99 with frame-of-reference conversion, and report its conversion status and errors from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
IDs	Required	The fully qualified record IDs to fetch, each in the format {partition}:{object-type}:{unique-id}. Specify up to 20 IDs without --attributes, or up to 100 with it.
Endpoint	Required	The service endpoint, for example https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, for example contoso-dp1.
Attributes	Optional	Dotted-path fields to return instead of whole records, for example data.Name and data.FileSourceInfo. Use it only when specific fields are requested; it can't be combined with --frame-of-reference.
Frame of reference	Optional	Convert measurements to SI, coordinates to WGS84, and dates to UTC on the server. The response then includes conversionStatuses, which is empty for records without measured or spatial fields.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Storage record: Get

Gets one ADME/OSDU record by record ID and returns the record content.

Azure MCP CLI

Example CLI command

azmcp adme storage record get --endpoint <endpoint> \
                                --data-partition <data-partition> \
                                --id <record-id> \
                                [--version <version>] \
                                [--attributes <path> [<path>...]]

Parameter	Type	Required	Description
id	string	Yes	The fully-qualified record ID {partition}:{object-type}:{unique-id}, such as opendes:well:W-99. Pass it verbatim as returned by azmcp adme storage record list.
endpoint	string	Yes	The service endpoint, such as https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, such as contoso-dp1.
version	64-bit integer	No	The numeric record version to retrieve, such as 1704779151123456. Omit to get the latest version. Use azmcp adme storage record version list to discover valid versions.
attributes	array of strings	No	Dotted-path fields to return from the requested record instead of the whole record, such as data.WellID and data.Name. Omit to return the full record.

MCP Server

Example prompts include:

• "Get the OSDU record opendes:well:W-99 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Get version 1704779151123456 of OSDU record opendes:well:W-99 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Get only data.WellID and data.Name from OSDU record opendes:well:W-99 at endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
ID	Required	The fully-qualified record ID {partition}:{object-type}:{unique-id}, for example opendes:well:W-99. Pass it verbatim as returned by azmcp adme storage record list.
Endpoint	Required	The service endpoint, such as https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, such as contoso-dp1.
Version	Optional	The numeric record version to retrieve, for example 1704779151123456. Omit to get the latest version. Use azmcp adme storage record version list to discover valid versions.
Attributes	Optional	Dotted-path fields to return from the requested record instead of the whole record, for example data.WellID and data.Name. Omit to return the full record.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Storage record: List

List multiple ADME and OSDU record IDs for a specified kind. This command returns only IDs and supports pagination.

Azure MCP CLI

Example CLI command

azmcp adme storage record list --endpoint <endpoint> \
                                 --data-partition <data-partition> \
                                 --kind <authority:source:entity-type:version> \
                                 [--limit <limit>] \
                                 [--cursor <cursor>]

Parameter	Type	Required	Description
kind	string	Yes	The fully qualified kind {authority}:{source}:{entityType}:{version}, such as osdu:wks:master-data--Well:1.0.0. Use azmcp adme schema list to discover valid kinds.
endpoint	string	Yes	The service endpoint, for example https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, for example contoso-dp1.
limit	integer	No	The number of record IDs to return in one page, from 1 through 100. Defaults to 10.
cursor	string	No	The cursor returned by a previous page. Omit for the first page. A null cursor in the response means there are no more pages.

MCP Server

Example prompts include:

• "List record IDs for kind osdu:wks:master-data--Well:1.0.0 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "List the first 25 record IDs for kind osdu:wks:master-data--Well:1.0.0 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."
• "Continue listing record IDs for kind osdu:wks:master-data--Well:1.0.0 by using the cursor from the previous response at endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
Kind	Required	The fully qualified kind {authority}:{source}:{entityType}:{version}, such as osdu:wks:master-data--Well:1.0.0. Use azmcp adme schema list to find valid kinds.
Endpoint	Required	The service endpoint, such as https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, such as contoso-dp1.
Limit	Optional	The number of record IDs to return in one page, from 1 through 100. Defaults to 10.
Cursor	Optional	The cursor returned by a previous page. Omit for the first page. A null cursor in the response means there are no more pages.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Storage record version: List

Lists versions of a single ADME/OSDU record by ID. Returns record versions ordered oldest first. Filters results to the specified Endpoint and Data partition.

Azure MCP CLI

Example CLI command

azmcp adme storage record version list --endpoint <endpoint> \
                                         --data-partition <data-partition> \
                                         --id <record-id> 

Parameter	Type	Required	Description
id	string	Yes	The fully qualified record ID {partition}:{object-type}:{unique-id}, such as opendes:well:W-99.
endpoint	string	Yes	The service endpoint, for example https://contoso.energy.azure.com.
data-partition	string	Yes	The data partition to target, for example contoso-dp1.

MCP Server

Example prompts include:

• "List all versions of record opendes:well:W-99 from endpoint https://contoso.energy.azure.com in data partition contoso-dp1."

Parameter	Required or optional	Description
ID	Required	The fully qualified record ID {partition}:{object-type}:{unique-id}, such as opendes:well:W-99.
Endpoint	Required	The service endpoint, for example https://contoso.energy.azure.com.
Data partition	Required	The data partition to target, for example contoso-dp1.

Tool annotation hints:

Destructive	Idempotent	Open World	Read Only	Secret	Local Required
❌	✅	❌	✅	❌	❌

Related content

• What are the Azure MCP Server tools?
• Get started using Azure MCP Server
• Azure Data Manager for Energy