Adopt
Vendor connectors and AI agents
New This page is new since 1.2.0-rc.1 (2026-11-06). Everything that changed
The products that analysts rank for data catalogues, metadata management and data governance, and the platforms that carry catalogues inside them. For each one: where its code and metadata model are published, how its fields map to DCAT-AP-UK, an add-on kit for the properties it lacks, and what it offers AI agents through the Model Context Protocol. Live connectors need a tenant and its credentials, so the portal ships kits and conversions rather than calling any vendor's service.
Nothing here is a test, an endorsement or a ranking by this project, and no kit has been tried against a live tenant. Positions in paywalled analyst reports come from the vendors' own announcements, labelled S, or from secondary sources, labelled K. Suppliers can correct an entry with the claim form on the integrations page.
Find your product
Filter by kind, report or what you need; select a product to see how to connect it.
| Product | Kind | Runs as K | Reports | Public repository | Field mapping | Add-on kit | Weaver | MCP server | In Claude |
|---|---|---|---|---|---|---|---|---|---|
| Atlan | Ranked | SaaS | 4 | Yes | Yes | Yes | Yes | Yes | Yes |
| Collibra Platform | Ranked | SaaS, On-prem | 3 | Yes | Yes | Yes | Yes | Yes | Yes |
| Alation Data Intelligence Platform | Ranked | SaaS, On-prem | 3 | Yes | No | No | No | Yes | No |
| Informatica Cloud Data Governance and Catalog (CDGC) | Ranked | SaaS | 4 | No | No | No | No | Yes | Yes |
| IBM watsonx.data intelligence | Ranked | SaaS, On-prem | 3 | Yes | No | No | No | Yes | No |
| data.world | Ranked | SaaS | 1 | Yes | Yes | Yes | Yes | No | No |
| OvalEdge | Ranked | SaaS, On-prem | 1 | Yes | No | No | No | Yes | Yes |
| Ataccama ONE | Ranked | SaaS, On-prem | 2 | No | No | No | No | Yes | No |
| DataGalaxy | Ranked | SaaS | 1 | No | No | No | No | Yes | No |
| Microsoft Purview (Data Map and Unified Catalog) | Platform | Cloud | 0 | Yes | Yes | Yes | Yes | No | Yes |
| Apache Atlas | Platform | Self-hosted, On-prem | 0 | Yes | Yes | Yes | Yes | No | No |
| OpenMetadata (and Collate) | Platform | Self-hosted, SaaS | 0 | Yes | Yes | Yes | Yes | Yes | No |
| DataHub (and DataHub Cloud) | Platform | Self-hosted, SaaS | 0 | Yes | Yes | Yes | Yes | Yes | Yes |
| Unity Catalog (open source and Databricks) | Platform | Self-hosted, Cloud | 0 | Yes | Yes | Yes | Yes | No | Yes |
| Google Cloud Dataplex Universal Catalog | Platform | Cloud | 0 | Yes | Yes | Yes | Yes | Yes | Yes |
| Snowflake (Horizon Catalog and Snowflake Open Catalog) | Platform | SaaS | 0 | Yes | Yes | Yes | Yes | Yes | Yes |
| AWS Glue Data Catalog | Platform | Cloud | 0 | Yes | Yes | Yes | Yes | Yes | Yes |
| Amazon SageMaker Catalog (Amazon DataZone) | Platform | Cloud | 0 | Yes | Yes | Yes | Yes | No | No |
| Oracle Cloud Infrastructure Data Catalog | Platform | Cloud | 0 | Yes | Yes | Yes | Yes | No | No |
| SAP Datasphere | Platform | SaaS | 0 | Yes | No | No | No | Yes | No |
| Cloudera (SDX) | Platform | On-prem, Cloud | 0 | Yes | Yes | Yes | Yes | No | No |
| Teradata Vantage | Platform | On-prem, Cloud | 0 | Yes | No | No | No | Yes | No |
| Qlik Talend Cloud (Talend Data Catalog) | Catalogue | SaaS, On-prem | 0 | No | No | No | No | Yes | Yes |
| Precisely Data Integrity Suite | Catalogue | SaaS | 0 | No | No | No | No | No | No |
| Select Star | Catalogue | SaaS | 0 | No | No | No | No | Yes | No |
| Secoda | Catalogue | SaaS, On-prem | 0 | Yes | No | No | No | Yes | No |
| Actian Data Intelligence Platform (formerly Zeenea) | Catalogue | SaaS | 0 | No | No | No | No | Yes | No |
| Apache Polaris (Iceberg REST catalog) | Open source | Self-hosted | 0 | Yes | Yes | Yes | Yes | No | No |
| OpenLineage (and Marquez) | Open source | Self-hosted | 0 | Yes | Yes | Yes | Yes | No | No |
| Amundsen | Open source | Self-hosted | 0 | Yes | Yes | No | No | No | No |
| Egeria (open metadata) | Open source | Self-hosted | 0 | Yes | No | No | No | No | No |
How it fits together
- REPORT an analyst report
- MAPPED a product with a field mapping
- PRODUCT · ROUTE another product, or a way to connect
- named in, or connects by
- connects AI agents by
Select a report, product or route to see its details. Without JavaScript, each card goes to its section.
The map as a list
- Metadata Management Solutions links to Atlan
- Data and Analytics Governance Platforms (first edition) links to Atlan
- Data Governance Solutions, Q3 2025 links to Atlan
- Enterprise Data Catalogs, Q3 2024 links to Atlan
- Atlan links to Add-on kit
- Atlan links to Weaver conversion
- Atlan uses MCP server for AI agents
- Atlan uses In Claude
- Data and Analytics Governance Platforms (first edition) links to Collibra Platform
- Data Governance Solutions, Q3 2025 links to Collibra Platform
- Enterprise Data Catalogs, Q3 2024 links to Collibra Platform
- Collibra Platform links to Add-on kit
- Collibra Platform links to Weaver conversion
- Collibra Platform uses MCP server for AI agents
- Collibra Platform uses In Claude
- Metadata Management Solutions links to Alation Data Intelligence Platform
- Data and Analytics Governance Platforms (first edition) links to Alation Data Intelligence Platform
- Data Governance Solutions, Q3 2025 links to Alation Data Intelligence Platform
- Alation Data Intelligence Platform uses MCP server for AI agents
- Metadata Management Solutions links to Informatica Cloud Data Governance and Catalog (CDGC)
- Data and Analytics Governance Platforms (first edition) links to Informatica Cloud Data Governance and Catalog (CDGC)
- Data Governance Solutions, Q3 2025 links to Informatica Cloud Data Governance and Catalog (CDGC)
- Enterprise Data Catalogs, Q3 2024 links to Informatica Cloud Data Governance and Catalog (CDGC)
- Informatica Cloud Data Governance and Catalog (CDGC) uses MCP server for AI agents
- Informatica Cloud Data Governance and Catalog (CDGC) uses In Claude
- Metadata Management Solutions links to IBM watsonx.data intelligence
- Data and Analytics Governance Platforms (first edition) links to IBM watsonx.data intelligence
- Data and Analytics Governance Platforms (second edition) links to IBM watsonx.data intelligence
- IBM watsonx.data intelligence uses MCP server for AI agents
- Data and Analytics Governance Platforms (first edition) links to data.world
- data.world links to Add-on kit
- data.world links to Weaver conversion
- Data and Analytics Governance Platforms (first edition) links to OvalEdge
- OvalEdge uses MCP server for AI agents
- OvalEdge uses In Claude
- Metadata Management Solutions links to Ataccama ONE
- Data and Analytics Governance Platforms (first edition) links to Ataccama ONE
- Ataccama ONE uses MCP server for AI agents
- Data and Analytics Governance Platforms (first edition) links to DataGalaxy
- DataGalaxy uses MCP server for AI agents
- Microsoft Purview (Data Map and Unified Catalog) links to Add-on kit
- Microsoft Purview (Data Map and Unified Catalog) links to Weaver conversion
- Microsoft Purview (Data Map and Unified Catalog) uses In Claude
- Apache Atlas links to Add-on kit
- Apache Atlas links to Weaver conversion
- OpenMetadata (and Collate) links to Add-on kit
- OpenMetadata (and Collate) links to Weaver conversion
- OpenMetadata (and Collate) uses MCP server for AI agents
- DataHub (and DataHub Cloud) links to Add-on kit
- DataHub (and DataHub Cloud) links to Weaver conversion
- DataHub (and DataHub Cloud) uses MCP server for AI agents
- DataHub (and DataHub Cloud) uses In Claude
- Unity Catalog (open source and Databricks) links to Add-on kit
- Unity Catalog (open source and Databricks) links to Weaver conversion
- Unity Catalog (open source and Databricks) uses In Claude
- Google Cloud Dataplex Universal Catalog links to Add-on kit
- Google Cloud Dataplex Universal Catalog links to Weaver conversion
- Google Cloud Dataplex Universal Catalog uses MCP server for AI agents
- Google Cloud Dataplex Universal Catalog uses In Claude
- Snowflake (Horizon Catalog and Snowflake Open Catalog) links to Add-on kit
- Snowflake (Horizon Catalog and Snowflake Open Catalog) links to Weaver conversion
- Snowflake (Horizon Catalog and Snowflake Open Catalog) uses MCP server for AI agents
- Snowflake (Horizon Catalog and Snowflake Open Catalog) uses In Claude
- AWS Glue Data Catalog links to Add-on kit
- AWS Glue Data Catalog links to Weaver conversion
- AWS Glue Data Catalog uses MCP server for AI agents
- AWS Glue Data Catalog uses In Claude
- Amazon SageMaker Catalog (Amazon DataZone) links to Add-on kit
- Amazon SageMaker Catalog (Amazon DataZone) links to Weaver conversion
- Oracle Cloud Infrastructure Data Catalog links to Add-on kit
- Oracle Cloud Infrastructure Data Catalog links to Weaver conversion
- SAP Datasphere uses MCP server for AI agents
- Cloudera (SDX) links to Add-on kit
- Cloudera (SDX) links to Weaver conversion
- Teradata Vantage uses MCP server for AI agents
- Qlik Talend Cloud (Talend Data Catalog) uses MCP server for AI agents
- Qlik Talend Cloud (Talend Data Catalog) uses In Claude
- Select Star uses MCP server for AI agents
- Secoda uses MCP server for AI agents
- Actian Data Intelligence Platform (formerly Zeenea) uses MCP server for AI agents
- Apache Polaris (Iceberg REST catalog) links to Add-on kit
- Apache Polaris (Iceberg REST catalog) links to Weaver conversion
- OpenLineage (and Marquez) links to Add-on kit
- OpenLineage (and Marquez) links to Weaver conversion
Products and how to connect them
Each product has steps in order. A field mapping cites the public file it was mapped from and the commit read; a product whose model is not published has no mapping, rather than a guessed one.
Named in the analyst reports (9)
Atlan
Mapped Kit Weaver MCP In Claude Runs as: SaaS S. Single-tenant SaaS on AWS, Azure or Google Cloud; a self-deployed runtime runs connectors inside the customer's network. (source, checked 14 November 2026.) In: Gartner metadata management 2025, Gartner D&A governance 2025, Forrester data governance 2025, Forrester data catalogues 2024
How to connect
- Add the DCAT-AP-UK properties to Atlan: load Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan). Custom metadata, sent to the typedef API as businessMetadataDefs (pyatlan/model/typedef.py). Atlan addresses custom metadata by internal names in its API, so map the kit's names to them when loading.
- Convert records both ways in the metadata weaver (choose Apache Atlas (Microsoft Purview, Atlan)), or in a pipeline:
node integrations/cli/validate.js --from atlas export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: Hosted MCP server at mcp.atlan.com/mcp with OAuth; the local server in the repository is deprecated and in maintenance only.
- In Claude: Atlan (plugin). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from pyatlan/model/assets/core/asset.py
Mapped from atlanhq/atlan-python. Custom metadata, sent to the typedef API as businessMetadataDefs (pyatlan/model/typedef.py). Atlan addresses custom metadata by internal names in its API, so map the kit's names to them when loading.
| DCAT-AP-UK | Atlan field | Note |
|---|---|---|
| title | name | |
| description | userDescription | Atlan keeps a crawled description and a user description; write the curated text to userDescription. |
| identifier | qualifiedName | Atlan builds qualifiedName from the connection; keep the DCAT-AP-UK identifier in custom metadata as well. |
| landingPage | sourceURL | |
| contactPoint | (none) | ownerUsers and ownerGroups name people and groups. Give a team contact form in custom metadata instead. |
| Every other property | businessAttributes.dcat_ap_uk | The extension place the add-on kit defines. |
Crosswalk as CSV Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan)
Public repositories (3)
- atlanhq/atlan-python: Python SDK; its typedef module defines custom metadata (commit
3c9c37ada6c5, checked 2026-11-07) - atlanhq/atlan-java: Java SDK (commit
b44068aa921d, checked 2026-11-07) - atlanhq/agent-toolkit: MCP server (commit
17b7ae3b4852, checked 2026-11-07)
MCP tools include search_assets, get_assets_by_dsl, traverse_lineage, update_assets, query_asset, create_dq_rules, schedule_dq_rules, update_dq_rules, delete_dq_rules.
Collibra Platform
Mapped Kit Weaver MCP In Claude Runs as: SaaS, On premises S. Collibra Platform in the cloud, or Collibra Platform Self-Hosted on infrastructure the customer chooses. (source, checked 14 November 2026.) In: Gartner D&A governance 2025, Forrester data governance 2025, Forrester data catalogues 2024
How to connect
- Add the DCAT-AP-UK properties to Collibra Platform: load Import API template and attribute types (Collibra). Attribute types and asset types defined in the operating model by an administrator; the Import API loads data, not the model itself.
- Convert records both ways in the metadata weaver (choose Collibra (Import API)), or in a pipeline:
node integrations/cli/validate.js --from collibra export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: A hosted MCP server on every instance at /rest/mcp with per-user OAuth, and an open-source local server (collibra/chip) using basic credentials.
- In Claude: Collibra (connector). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from input-file-examples/src/test/resources (asset JSON with attributes keyed by attribute type name)
Mapped from collibra/import-api-examples. Attribute types and asset types defined in the operating model by an administrator; the Import API loads data, not the model itself.
| DCAT-AP-UK | Collibra Platform field | Note |
|---|---|---|
| title | name | Collibra asset name; the identifier block also names the domain and community. |
| description | attributes["Description"] | Single values are written as [{"value": ...}], as in the repository examples. |
| identifier | identifier.externalEntityId | With identifier.externalSystemId naming the source catalogue. |
| keyword | tags | The repository has an example that adds assets with tags. |
| contactPoint | (none) | Responsibilities name people. Give a team contact form as an attribute. |
| Every other property | attributes["DCAT-AP-UK <property>"] | The extension place the add-on kit defines. |
Crosswalk as CSV Import API template and attribute types (Collibra)
Public repositories (3)
- collibra/import-api-examples: Import API examples and input files (commit
48e1e29c2a50, checked 2026-11-07) - collibra/custom-technical-lineage-examples: Custom technical lineage files (commit
be0662134504, checked 2026-11-07) - collibra/chip: Local MCP server (Go) with embedded agent skills (commit
05d201f35d20, checked 2026-11-07)
MCP tools include discover_data_assets, discover_business_glossary, search_asset_keyword, get_asset_details, get_lineage_upstream, get_lineage_downstream, create_asset, edit_asset, list_data_contracts, pull_data_contract_manifest, push_data_contract_manifest, create_dq_rule.
Alation Data Intelligence Platform
MCP Runs as: SaaS, On premises S. Alation Cloud Service, or customer-managed Alation. (source, checked 14 November 2026.) In: Gartner metadata management 2025, Gartner D&A governance 2025, Forrester data governance 2025
How to connect
- No field mapping yet: No public repository publishes Alation's catalogue object model, so no field mapping is given rather than a guessed one. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: An MCP server (alation-ai-agent-mcp) in the agent SDK, over stdio or HTTP.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- Alation/alation-ai-agent-sdk: AI agent SDK (commit
ceac7b6ad6ed, checked 2026-11-07)
MCP tools include alation_context, bulk_retrieval, get_data_products, get_custom_fields_definitions, get_data_dictionary_instructions, data_quality_tool, get_data_sources_tool.
Informatica Cloud Data Governance and Catalog (CDGC)
MCP In Claude Runs as: SaaS S. Cloud Data Governance and Catalog is a service of the Intelligent Data Management Cloud; its scanners can run in the customer network. The older on-premises Enterprise Data Catalog and Axon are a separate product line, not counted here. (source, checked 14 November 2026.) In: Gartner metadata management 2025, Gartner D&A governance 2025, Forrester data governance 2025, Forrester data catalogues 2024
How to connect
- No field mapping yet: No public repository of CDGC's model was found. No field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: Informatica MCP servers over streamable HTTP (MCP protocol 2025-06-18), with session ID or JSON web token authentication.
- In Claude: Informatica (informatica-for-claude-platform) (plugin); Informatica Catalog Discovery (connector); Informatica Data Exploration (connector). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
MCP tools include search_metadata, get_asset_details.
IBM watsonx.data intelligence
MCP Runs as: SaaS, On premises S. Managed on IBM Cloud, or Cloud Pak for Data on a private, on-premises cluster. (source, checked 14 November 2026.) In: Gartner metadata management 2025, Gartner D&A governance 2025, Gartner D&A governance 2026
How to connect
- No field mapping yet: The custom asset type payload is documented but not published as a schema file, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: A local MCP server (ibm-watsonx-data-intelligence-mcp-server, stdio or HTTP) and, from 2026, a managed MCP server.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- IBM/data-intelligence-mcp-server: MCP server with agent skills (Apache 2.0) (commit
ba81d268216c, checked 2026-11-07)
MCP tools include get_asset_details, list_business_terms, get_lineage_graph, get_data_quality, create_data_quality_rule, create_or_update_metadata_import, create_glossary_from_files, get_data_product_details, bulk_publish_assets.
data.world
Mapped Kit Weaver Runs as: SaaS S. Private instance on a multi-tenant platform, or a single-tenant installation; both run by data.world. (source, checked 14 November 2026.) In: Gartner D&A governance 2025
How to connect
- Add the DCAT-AP-UK properties to data.world: load Dataset properties template (data.world). Free-form dataset properties.
- Convert records both ways in the metadata weaver (choose data.world), or in a pipeline:
node integrations/cli/validate.js --from dataworld export.jsonchecks an export as DCAT-AP-UK. - Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from datadotworld/client/_swagger/models/dataset_create_request.py
Mapped from datadotworld/data.world-py. Free-form dataset properties.
| DCAT-AP-UK | data.world field | Note |
|---|---|---|
| title | title | |
| description | description | |
| abstract | summary | |
| keyword | tags | |
| licence | license | The API accepts a fixed list (CC-BY, ODC-BY, Public Domain, Other and others). OGL v3.0 is not on it: send Other and keep the licence address in properties. |
| Every other property | properties | The extension place the add-on kit defines. |
Public repositories (1)
- datadotworld/data.world-py: Python client with the dataset API models (commit
b1c03aa590ff, checked 2026-11-07)
OvalEdge
MCP In Claude Runs as: SaaS, On premises S. On premises, in the customer's own cloud account, or as SaaS. (source, checked 14 November 2026.) In: Gartner D&A governance 2025
How to connect
- No field mapping yet: The repository wraps OvalEdge's REST API but does not publish its object model as a schema, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: An MCP server that runs locally over stdio or remotely over HTTPS at /mcp, acting with the signed-in user's permissions; writes ask the user to confirm first.
- In Claude: claude-ovaledge (plugin). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- ovaledge/oe_mcp: Open-source MCP server (MIT) with a Claude Code plugin and a Codex plugin (commit
9587d7c18075, checked 2026-11-07)
MCP tools include asset_explorer, asset_details, asset_lineage, metadata_changes_between_crawls, update_asset_descriptions, update_cde_associations, access_explorer, create_glossary_term, create_tag, create_service_request, update_governance_roles, update_custom_field_value, dq_rule_advisor, dq_rule_manager, knowledge_search.
Ataccama ONE
MCP Runs as: SaaS, On premises S. Ataccama Cloud (SaaS, optionally hybrid), or self-managed on premises. (source, checked 14 November 2026.) In: Gartner metadata management 2025, Gartner D&A governance 2025
How to connect
- No field mapping yet: No public repository of the metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: The Data Quality and Catalog MCP server, at your Ataccama ONE host followed by /mcp (from version 17.2.0), signed in with Ataccama ONE credentials. Ataccama Cloud only, enabled on request; self-managed deployments are not supported.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
DataGalaxy
MCP Runs as: SaaS S. The vendor's own marketplace listing gives delivery as Software as a Service. (source, checked 14 November 2026.) In: Gartner D&A governance 2025
How to connect
- No field mapping yet: No public repository of the metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: A read-only MCP server that searches objects, links, users and tags, returns ownership, definitions and lineage, and keeps DataGalaxy's roles and access controls. Its address is shown in each DataGalaxy instance, with a personal token.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
Cloud and data platforms (13)
Microsoft Purview (Data Map and Unified Catalog)
Mapped Kit Weaver In Claude Runs as: In a cloud platform S. A Microsoft cloud service; Microsoft calls the catalog a single, integrated SaaS framework. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Microsoft Purview (Data Map and Unified Catalog): load Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan). Custom types (typedefs) and business metadata through the Atlas API.
- Convert records both ways in the metadata weaver (choose Apache Atlas (Microsoft Purview, Atlan)), or in a pipeline:
node integrations/cli/validate.js --from atlas export.jsonchecks an export as DCAT-AP-UK. - In Claude: Microsoft Learn (connector); microsoft-docs (plugin). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from addons/models/0000-Area0/0010-base_model.json
Mapped from apache/atlas. Custom types (typedefs) and business metadata through the Atlas API.
| DCAT-AP-UK | Microsoft Purview (Data Map and Unified Catalog) field | Note |
|---|---|---|
| title | name | Asset.name in the Atlas base model. |
| description | userDescription | Asset also has description, filled by scans. |
| identifier | qualifiedName | Referenceable.qualifiedName is the unique key; keep the DCAT-AP-UK identifier in business metadata as well. |
| contactPoint | (none) | Asset.owner names a person. Give a team contact form in business metadata. |
| Every other property | businessAttributes.dcat_ap_uk | The extension place the add-on kit defines. |
Crosswalk as CSV Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan)
Public repositories (3)
- apache/atlas: The type system Purview's Data Map uses (commit
8b3f674e0f25, checked 2026-11-07) - microsoft/purview-custom-connector-solution-accelerator: Custom connector accelerator using Atlas JSON (commit
7a8eb4af1e9b, checked 2026-11-07) - microsoft/Purview-Custom-Types-Tool-Solution-Accelerator: Tool for building custom types (commit
544358eab84c, checked 2026-11-07)
Apache Atlas
Mapped Kit Weaver Runs as: Open source, self-hosted, On premises V. Apache-licensed source in the pinned repository; the customer runs it. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Apache Atlas: load Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan). Typedefs and business metadata.
- Convert records both ways in the metadata weaver (choose Apache Atlas (Microsoft Purview, Atlan)), or in a pipeline:
node integrations/cli/validate.js --from atlas export.jsonchecks an export as DCAT-AP-UK. - Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from addons/models/0000-Area0/0010-base_model.json
Mapped from apache/atlas. Typedefs and business metadata.
| DCAT-AP-UK | Apache Atlas field | Note |
|---|---|---|
| title | name | |
| description | description | |
| identifier | qualifiedName | |
| Every other property | businessAttributes.dcat_ap_uk | The extension place the add-on kit defines. |
Crosswalk as CSV Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan)
Public repositories (1)
- apache/atlas: Source and base type model (commit
8b3f674e0f25, checked 2026-11-07)
OpenMetadata (and Collate)
Mapped Schema-checked Kit Weaver MCP Runs as: Open source, self-hosted, SaaS S. Self-hosted with Docker Compose or Kubernetes, or Collate, the managed SaaS. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to OpenMetadata (and Collate): load Custom properties for the table entity (OpenMetadata). Custom properties on an entity type (type/customProperty.json), stored in the entity's extension object.
- Convert records both ways in the metadata weaver (choose OpenMetadata), or in a pipeline:
node integrations/cli/validate.js --from openmetadata export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: A built-in MCP server at /mcp on every OpenMetadata server since 1.8.0.
- What the portal produces for OpenMetadata (and Collate) is checked on every build against OpenMetadata's table schema, every worked example; the id and columns are added at load time.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from openmetadata-spec/src/main/resources/json/schema/entity/data/table.json
Mapped from open-metadata/OpenMetadata. Custom properties on an entity type (type/customProperty.json), stored in the entity's extension object.
| DCAT-AP-UK | OpenMetadata (and Collate) field | Note |
|---|---|---|
| title | displayName | name is a machine name; use displayName for the title. |
| description | description | |
| keyword | tags[].tagFQN | Tags are classification terms; free keywords need a classification of their own. |
| landingPage | sourceUrl | |
| contactPoint | (none) | owners names users and teams. Give a team contact form as a custom property. |
| Every other property | extension | The extension place the add-on kit defines. |
Crosswalk as CSV Custom properties for the table entity (OpenMetadata)
Public repositories (2)
- open-metadata/OpenMetadata: Source and JSON Schema entity model (commit
9abf4c712ca2, checked 2026-11-07) - open-metadata/ai-sdk: AI SDK over the built-in MCP server (commit
01d572f5c427, checked 2026-11-07)
MCP tools include search_metadata.
DataHub (and DataHub Cloud)
Mapped Schema-checked Kit Weaver MCP In Claude Runs as: Open source, self-hosted, SaaS S. Self-hosted DataHub Core, or the managed service now called DataHub Cloud (previously Acryl Cloud). (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to DataHub (and DataHub Cloud): load Structured properties (DataHub). Structured properties (metadata-ingestion/examples/structured_properties), and free customProperties.
- Convert records both ways in the metadata weaver (choose DataHub), or in a pipeline:
node integrations/cli/validate.js --from datahub export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: An MCP server over stdio or HTTP, with search, lineage and editing tools.
- In Claude: DataHub (connector); DataHub Cloud (plugin); datahub-skills (plugin). See In Claude.
- What the portal produces for DataHub (and DataHub Cloud) is checked on every build against the DataHub project's own structured property examples (the kit's keys and types).
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from metadata-models/src/main/pegasus/com/linkedin/dataset/DatasetProperties.pdl
Mapped from datahub-project/datahub. Structured properties (metadata-ingestion/examples/structured_properties), and free customProperties.
| DCAT-AP-UK | DataHub (and DataHub Cloud) field | Note |
|---|---|---|
| title | name | |
| description | description | |
| keyword | tags | |
| landingPage | externalUrl | From the ExternalReference record DatasetProperties includes. |
| issued | created.time | Milliseconds since 1970; written as a date. |
| modified | lastModified.time | |
| Every other property | customProperties | The extension place the add-on kit defines. |
Public repositories (2)
- datahub-project/datahub: Source and PDL metadata model (commit
f65d0be5605c, checked 2026-11-07) - acryldata/mcp-server-datahub: Official MCP server (commit
e4de4c021eba, checked 2026-11-07)
MCP tools include search, get_entities, get_lineage, get_lineage_paths_between, list_schema_fields, get_dataset_queries, update_description, add_tags, add_terms, add_owners, set_domains, add_structured_properties.
Unity Catalog (open source and Databricks)
Mapped Kit Weaver In Claude Runs as: Open source, self-hosted, In a cloud platform S. Built into Databricks on AWS, Azure and Google Cloud; the open-source implementation is in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Unity Catalog (open source and Databricks): load Table properties (Unity Catalog). The table's properties map of strings.
- Convert records both ways in the metadata weaver (choose Unity Catalog), or in a pipeline:
node integrations/cli/validate.js --from unity export.jsonchecks an export as DCAT-AP-UK. - In Claude: databricks (plugin); Databricks Genie (connector). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from api/all.yaml (TableInfo)
Mapped from unitycatalog/unitycatalog. The table's properties map of strings.
| DCAT-AP-UK | Unity Catalog (open source and Databricks) field | Note |
|---|---|---|
| description | comment | |
| issued | created_at | Milliseconds since 1970. |
| modified | updated_at | |
| title | (none) | name is a three-part SQL name, not a title. Keep the title in properties. |
| Every other property | properties | The extension place the add-on kit defines. |
Public repositories (1)
- unitycatalog/unitycatalog: Open-source server and OpenAPI model (commit
2a2c4106ec6b, checked 2026-11-07)
Google Cloud Dataplex Universal Catalog
Mapped Kit Weaver MCP In Claude Runs as: In a cloud platform V. A Google Cloud service: its API is defined in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Google Cloud Dataplex Universal Catalog: load Aspect type (Dataplex Universal Catalog). Aspect types with a metadata template; an entry carries aspects keyed by aspect type.
- Convert records both ways in the metadata weaver (choose Dataplex Universal Catalog), or in a pipeline:
node integrations/cli/validate.js --from dataplex export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: MCP Toolbox for Databases has a prebuilt Dataplex tool set.
- In Claude: Google Cloud BigQuery (connector). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from google/cloud/dataplex/v1/catalog.proto (Entry
Mapped from googleapis/googleapis. Aspect types with a metadata template; an entry carries aspects keyed by aspect type.
| DCAT-AP-UK | Google Cloud Dataplex Universal Catalog field | Note |
|---|---|---|
| title | entrySource.displayName | |
| description | entrySource.description | |
| issued | entrySource.createTime | |
| modified | entrySource.updateTime | |
| keyword | entrySource.labels | Labels are key and value pairs; each keyword becomes a key with an empty value. |
| Every other property | aspects["dcat-ap-uk"].data | The extension place the add-on kit defines. |
Public repositories (2)
- googleapis/googleapis: The Dataplex v1 catalog API definitions (commit
bf54c1031efd, checked 2026-11-07) - googleapis/genai-toolbox: MCP Toolbox (commit
8d54da0d2c40, checked 2026-11-07)
MCP tools include search_entries, lookup_entry, lookup_context, search_dq_scans, list_data_products.
Snowflake (Horizon Catalog and Snowflake Open Catalog)
Mapped Schema-checked Kit Weaver MCP In Claude Runs as: SaaS S. Snowflake runs both catalogues as managed services. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Snowflake (Horizon Catalog and Snowflake Open Catalog): load Table properties (Apache Iceberg REST catalogs: Snowflake Open Catalog, Apache Polaris). Iceberg table properties (strings), kept with the table by Snowflake Open Catalog, which Snowflake documents as a managed service for Apache Polaris built on the Iceberg REST protocol (https://other-docs.snowflake.com/polaris/overview.html, checked 7 November 2026), and by any Iceberg REST catalog. Snowflake now points new customers to Horizon Catalog for Iceberg tables. Snowflake's own objects carry comments and tags; their model is not published as a schema.
- Convert records both ways in the metadata weaver (choose Apache Iceberg REST (Snowflake Open Catalog, Polaris)), or in a pipeline:
node integrations/cli/validate.js --from iceberg export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: The official Snowflake MCP Server, managed by Snowflake and built on Cortex Agents; the open-source Snowflake-Labs/mcp server is deprecated and points to it.
- In Claude: Snowflake (connector); snowflake-cortex-code (plugin). See In Claude.
- What the portal produces for Snowflake (Horizon Catalog and Snowflake Open Catalog) is checked on every build against the Iceberg REST specification (CreateTableRequest), every worked example; the table schema is added at load time.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from open-api/rest-catalog-open-api.yaml (CreateTableRequest)
Mapped from apache/iceberg. Iceberg table properties (strings), kept with the table by Snowflake Open Catalog, which Snowflake documents as a managed service for Apache Polaris built on the Iceberg REST protocol (https://other-docs.snowflake.com/polaris/overview.html, checked 7 November 2026), and by any Iceberg REST catalog. Snowflake now points new customers to Horizon Catalog for Iceberg tables. Snowflake's own objects carry comments and tags; their model is not published as a schema.
| DCAT-AP-UK | Snowflake (Horizon Catalog and Snowflake Open Catalog) field | Note |
|---|---|---|
| description | properties.comment | TableProperties.COMMENT in Apache Iceberg (core/src/main/java/org/apache/iceberg/TableProperties.java at the cited commit); Spark and Trino store CREATE TABLE ... COMMENT there. |
| title | (none) | name is the table identifier, not a title. Keep the title in properties. |
| Every other property | properties | The extension place the add-on kit defines. |
Crosswalk as CSV Table properties (Apache Iceberg REST catalogs: Snowflake Open Catalog, Apache Polaris)
Public repositories (3)
- apache/iceberg: The Iceberg REST catalog specification (open-api/rest-catalog-open-api.yaml) (commit
0cb5041b9c7a, checked 2026-11-07) - apache/polaris: Apache Polaris (commit
5f87d86e5f64, checked 2026-11-07) - Snowflake-Labs/mcp: Former community MCP server (commit
662cb486395d, checked 2026-11-07)
AWS Glue Data Catalog
Mapped Schema-checked Kit Weaver MCP In Claude Runs as: In a cloud platform V. An AWS service: its API is defined in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to AWS Glue Data Catalog: load Table parameters (AWS Glue Data Catalog). The table's Parameters map of strings.
- Convert records both ways in the metadata weaver (choose AWS Glue Data Catalog), or in a pipeline:
node integrations/cli/validate.js --from glue export.jsonchecks an export as DCAT-AP-UK. - Give AI agents the catalogue through its MCP server: The AWS Data Processing MCP server (awslabs.aws-dataprocessing-mcp-server), local over stdio with an AWS profile; changes need the --allow-write flag.
- In Claude: aws-data-analytics (plugin). See In Claude.
- What the portal produces for AWS Glue Data Catalog is checked on every build against AWS's own Glue model (TableInput in botocore), every worked example.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from botocore/data/glue/2017-03-31/service-2.json (TableInput)
Mapped from boto/botocore. The table's Parameters map of strings.
| DCAT-AP-UK | AWS Glue Data Catalog field | Note |
|---|---|---|
| description | Description | |
| title | (none) | Name is the table identifier, not a title. Keep the title in Parameters. |
| contactPoint | (none) | Owner names a person. Give a team contact form in Parameters. |
| Every other property | Parameters | The extension place the add-on kit defines. |
Public repositories (2)
- boto/botocore: AWS service models (commit
45d5dde95e31, checked 2026-11-07) - awslabs/mcp: AWS MCP servers (commit
d33705e0890e, checked 2026-11-07)
MCP tools include manage_aws_glue_databases, manage_aws_glue_tables, manage_aws_glue_partitions, manage_aws_glue_catalog, manage_aws_glue_connections, manage_aws_glue_crawlers.
Amazon SageMaker Catalog (Amazon DataZone)
Mapped Schema-checked Kit Weaver Runs as: In a cloud platform V. An AWS service (Amazon DataZone API): defined in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Amazon SageMaker Catalog (Amazon DataZone): load Metadata form type in Smithy (Amazon SageMaker Catalog, DataZone). Metadata form types, defined in Smithy and attached to an asset as forms.
- Convert records both ways in the metadata weaver (choose Amazon SageMaker Catalog (DataZone)), or in a pipeline:
node integrations/cli/validate.js --from datazone export.jsonchecks an export as DCAT-AP-UK. - What the portal produces for Amazon SageMaker Catalog (Amazon DataZone) is checked on every build against AWS's own DataZone model (CreateAssetInput in botocore), every worked example; domain, asset type and project are added at load time.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from botocore/data/datazone/2018-05-10/service-2.json (CreateAssetInput
Mapped from boto/botocore. Metadata form types, defined in Smithy and attached to an asset as forms.
| DCAT-AP-UK | Amazon SageMaker Catalog (Amazon DataZone) field | Note |
|---|---|---|
| title | name | |
| description | description | |
| identifier | externalIdentifier | |
| keyword | (none) | glossaryTerms holds glossary term identifiers, not free keywords. Keywords stay in the form. |
| Every other property | formsInput[DcatApUkForm].content | The extension place the add-on kit defines. |
Crosswalk as CSV Metadata form type in Smithy (Amazon SageMaker Catalog, DataZone)
Public repositories (1)
- boto/botocore: AWS service models (commit
45d5dde95e31, checked 2026-11-07)
Oracle Cloud Infrastructure Data Catalog
Mapped Schema-checked Kit Weaver Runs as: In a cloud platform V. An Oracle Cloud Infrastructure service: its client is in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Oracle Cloud Infrastructure Data Catalog: load Custom properties (OCI Data Catalog). Custom properties (create_custom_property_details.py), held on an entity as custom property members.
- Convert records both ways in the metadata weaver (choose OCI Data Catalog), or in a pipeline:
node integrations/cli/validate.js --from oci export.jsonchecks an export as DCAT-AP-UK. - What the portal produces for Oracle Cloud Infrastructure Data Catalog is checked on every build against the OCI SDK's CreateCustomPropertyDetails (the custom property kit).
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from src/oci/data_catalog/models/entity.py
Mapped from oracle/oci-python-sdk. Custom properties (create_custom_property_details.py), held on an entity as custom property members.
| DCAT-AP-UK | Oracle Cloud Infrastructure Data Catalog field | Note |
|---|---|---|
| title | displayName | |
| description | description | |
| Every other property | customPropertyMembers | The extension place the add-on kit defines. |
Public repositories (1)
- oracle/oci-python-sdk: OCI Python SDK (commit
ec1d21a37357, checked 2026-11-07)
SAP Datasphere
MCP Runs as: SaaS S. A data service built on SAP Business Technology Platform. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: SAP does not publish the catalogue's model as a schema in a public repository, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: Community MCP servers only, not published or supported by SAP. They read spaces and catalogue assets through Datasphere's Catalog and Consumption APIs.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- MarioDeFelipe/sap-datasphere-mcp: A community MCP server for Datasphere (commit
587811443b03, checked 2026-11-07)
Cloudera (SDX)
Mapped Kit Weaver Runs as: On premises, In a cloud platform S. Cloudera Data Catalog on premises (OpenShift or ECS clusters), and in Cloudera public cloud on AWS, Azure and Google Cloud. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Cloudera (SDX): load Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan). Atlas typedefs and business metadata. Cloudera's documentation names Apache Atlas as the metadata store behind its governance and Data Catalog (https://docs.cloudera.com/runtime/7.3.2/concepts-governance.html, checked 7 November 2026).
- Convert records both ways in the metadata weaver (choose Apache Atlas (Microsoft Purview, Atlan)), or in a pipeline:
node integrations/cli/validate.js --from atlas export.jsonchecks an export as DCAT-AP-UK. - Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from addons/models/0000-Area0/0010-base_model.json
Mapped from apache/atlas. Atlas typedefs and business metadata. Cloudera's documentation names Apache Atlas as the metadata store behind its governance and Data Catalog (https://docs.cloudera.com/runtime/7.3.2/concepts-governance.html, checked 7 November 2026).
| DCAT-AP-UK | Cloudera (SDX) field | Note |
|---|---|---|
| title | name | |
| description | userDescription | |
| Every other property | businessAttributes.dcat_ap_uk | The extension place the add-on kit defines. |
Crosswalk as CSV Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan)
Public repositories (1)
- apache/atlas: Apache Atlas (commit
8b3f674e0f25, checked 2026-11-07)
Teradata Vantage
MCP Runs as: On premises, In a cloud platform S. On premises, public cloud, hybrid and multi-cloud. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: No public catalogue model, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: An MCP server installed with uv or pip, connecting with a database URI; it reads schemas and column metadata and runs SQL.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- teradata/teradata-mcp-server: Teradata's MCP server (tools for querying (commit
2df7fcbcc47b, checked 2026-11-07)
Other catalogue products, SaaS and on-premises (5)
Qlik Talend Cloud (Talend Data Catalog)
MCP In Claude Runs as: SaaS, On premises S. On premises, cloud, or hybrid (cloud with on-premises harvesting). (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: No public metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: An MCP server in Qlik Talend Cloud, switched on per tenant by an administrator, giving AI assistants governed data sets and data products.
- In Claude: Qlik (connector). See In Claude.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
Precisely Data Integrity Suite
Runs as: SaaS S. Interoperable SaaS services sharing one catalogue; execution agents can run where the data is. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: No public metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
Select Star
MCP Runs as: SaaS S. Select Star describes itself as a SaaS provider hosted on AWS. No self-hosted option was found. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: No public metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: A hosted MCP server at mcp.production.selectstar.com/mcp, used with a Select Star API token.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
Secoda
MCP Runs as: SaaS, On premises S. SaaS, or self-hosted on Kubernetes (EKS, GKE, Helm) or Docker Compose. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: No public metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: A hosted MCP server in each Secoda workspace, and an open-source server that uses a Secoda API token.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- secoda/secoda-mcp: Secoda's MCP server (MIT) (commit
6cd32bf4e716, checked 2026-11-07)
MCP tools include search_data_assets, retrieve_entity, entity_lineage, search_documentation, run_sql.
Actian Data Intelligence Platform (formerly Zeenea)
MCP Runs as: SaaS S. Actian describes the platform as cloud-native SaaS. A March 2025 Actian press release also says it can be deployed across hybrid, cloud and on-premises environments; the product page is followed here until Actian documents an on-premises install. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: No public metadata model was found, so no field mapping is given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Give AI agents the catalogue through its MCP server: A managed MCP server connecting AI assistants such as Claude and Microsoft Copilot to the platform's metadata.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
No public repository found.
Open source and open standards (4)
Apache Polaris (Iceberg REST catalog)
Mapped Schema-checked Kit Weaver Runs as: Open source, self-hosted V. Apache-licensed source in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to Apache Polaris (Iceberg REST catalog): load Table properties (Apache Iceberg REST catalogs: Snowflake Open Catalog, Apache Polaris). Iceberg table properties.
- Convert records both ways in the metadata weaver (choose Apache Iceberg REST (Snowflake Open Catalog, Polaris)), or in a pipeline:
node integrations/cli/validate.js --from iceberg export.jsonchecks an export as DCAT-AP-UK. - What the portal produces for Apache Polaris (Iceberg REST catalog) is checked on every build against the Iceberg REST specification (CreateTableRequest), every worked example; the table schema is added at load time.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from open-api/rest-catalog-open-api.yaml (CreateTableRequest)
Mapped from apache/iceberg. Iceberg table properties.
| DCAT-AP-UK | Apache Polaris (Iceberg REST catalog) field | Note |
|---|---|---|
| description | properties.comment | TableProperties.COMMENT in Apache Iceberg (core/src/main/java/org/apache/iceberg/TableProperties.java at the cited commit). |
| Every other property | properties | The extension place the add-on kit defines. |
Crosswalk as CSV Table properties (Apache Iceberg REST catalogs: Snowflake Open Catalog, Apache Polaris)
Public repositories (2)
- apache/polaris: Source (commit
5f87d86e5f64, checked 2026-11-07) - apache/iceberg: The Iceberg REST catalog specification Polaris implements (commit
0cb5041b9c7a, checked 2026-11-07)
OpenLineage (and Marquez)
Mapped Schema-checked Kit Weaver Runs as: Open source, self-hosted V. An open specification and client libraries in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Add the DCAT-AP-UK properties to OpenLineage (and Marquez): load Custom dataset facet schema (OpenLineage). Custom dataset facets; the kit defines a dcat_ap_uk facet.
- Convert records both ways in the metadata weaver (choose OpenLineage dataset), or in a pipeline:
node integrations/cli/validate.js --from openlineage export.jsonchecks an export as DCAT-AP-UK. - What the portal produces for OpenLineage (and Marquez) is checked on every build against the OpenLineage specification (Dataset and the documentation facet), every worked example; the facet kit is a valid JSON Schema.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from spec/OpenLineage.json (Dataset) and spec/facets/DocumentationDatasetFacet.json
Mapped from OpenLineage/OpenLineage. Custom dataset facets; the kit defines a dcat_ap_uk facet.
| DCAT-AP-UK | OpenLineage (and Marquez) field | Note |
|---|---|---|
| description | facets.documentation.description | |
| identifier | namespace and name | A dataset is identified by its namespace and name; keep the DCAT-AP-UK identifier in the facet as well. |
| contactPoint | (none) | The ownership facet names people and teams. Give a team contact form in the facet. |
| Every other property | facets.dcat_ap_uk | The extension place the add-on kit defines. |
Public repositories (1)
- OpenLineage/OpenLineage: The OpenLineage specification and its dataset facets (commit
6e9fd6447831, checked 2026-11-07)
Amundsen
Mapped Runs as: Open source, self-hosted V. Apache-licensed source in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- Map the fields by hand with the field mapping below. Amundsen has no general extension place for extra properties; programmatic descriptions can carry text.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Field mapping, from databuilder/databuilder/models/table_metadata.py (TableMetadata)
Mapped from amundsen-io/amundsen. Amundsen has no general extension place for extra properties; programmatic descriptions can carry text.
| DCAT-AP-UK | Amundsen field | Note |
|---|---|---|
| description | description | |
| keyword | tags | |
| title | (none) | name is the table name, not a title. |
| Every other property | (none) | No extension place, so other properties are not carried. |
Public repositories (1)
- amundsen-io/amundsen: Source (commit
469c1b778340, checked 2026-11-07)
Egeria (open metadata)
Runs as: Open source, self-hosted V. Apache-licensed source in the pinned repository. (source, checked 14 November 2026.) In: Not ranked in these reports: a platform catalogue.
How to connect
- No field mapping yet: The open metadata types are large and spread across many definitions; a mapping is not yet given. Meanwhile, export DCAT-AP-UK records as JSON-LD or Turtle from the record builder and use the product's own DCAT or JSON import.
- Check every record against DCAT-AP-UK with the validator before it is published or harvested.
Public repositories (1)
- odpi/egeria: Source of the open metadata types and repository services (commit
9c2bb9761329, checked 2026-11-07)
AI agents, MCP servers and skills
What each vendor offers AI agents that find, explain and govern catalogue metadata. An agent can use these alongside this portal's own MCP server, which checks and converts DCAT-AP-UK records.
| Product | MCP server | Tools | Skills and plugins | Source |
|---|---|---|---|---|
| Atlan | Hosted MCP server at mcp.atlan.com/mcp with OAuth; the local server in the repository is deprecated and in maintenance only. | search_assets, get_assets_by_dsl, traverse_lineage, update_assets, query_asset, create_dq_rules, schedule_dq_rules, update_dq_rules, delete_dq_rules | An official Claude Code plugin, Codex and Cursor plugins, and agent skills in the skills folder (atlan-search). | Source V |
| Collibra Platform | A hosted MCP server on every instance at /rest/mcp with per-user OAuth, and an open-source local server (collibra/chip) using basic credentials. | discover_data_assets, discover_business_glossary, search_asset_keyword, get_asset_details, get_lineage_upstream, get_lineage_downstream, create_asset, edit_asset, list_data_contracts, pull_data_contract_manifest, push_data_contract_manifest, create_dq_rule | Skills embedded in the server and served through list_collibra_skills and load_collibra_skill (collibra/discovery, collibra/lineage, collibra/asset-create, collibra/asset-edit). | Source V |
| Alation Data Intelligence Platform | An MCP server (alation-ai-agent-mcp) in the agent SDK, over stdio or HTTP. | alation_context, bulk_retrieval, get_data_products, get_custom_fields_definitions, get_data_dictionary_instructions, data_quality_tool, get_data_sources_tool | A LangChain distribution of the same tools. | Source V |
| Informatica Cloud Data Governance and Catalog (CDGC) | Informatica MCP servers over streamable HTTP (MCP protocol 2025-06-18), with session ID or JSON web token authentication. | search_metadata, get_asset_details | Connection guides for Claude Desktop, Cursor, VS Code and Python, and an MCP tool in Informatica's own AI Agent Engineering service. | Source S |
| IBM watsonx.data intelligence | A local MCP server (ibm-watsonx-data-intelligence-mcp-server, stdio or HTTP) and, from 2026, a managed MCP server. | get_asset_details, list_business_terms, get_lineage_graph, get_data_quality, create_data_quality_rule, create_or_update_metadata_import, create_glossary_from_files, get_data_product_details, bulk_publish_assets | Agent skills shipped with the server, installed with wxdi-setup-skills (curate-glossary, lineage, data-quality, onboard-and-enrich, data-product-creation and others). | Source V |
| data.world | No MCP server was found in a public repository. | Not listed | K | |
| OvalEdge | An MCP server that runs locally over stdio or remotely over HTTPS at /mcp, acting with the signed-in user's permissions; writes ask the user to confirm first. | asset_explorer, asset_details, asset_lineage, metadata_changes_between_crawls, update_asset_descriptions, update_cde_associations, access_explorer, create_glossary_term, create_tag, create_service_request, update_governance_roles, update_custom_field_value, dq_rule_advisor, dq_rule_manager, knowledge_search | A Claude Code plugin in the repository (plugins/claude-ovaledge) with the ovaledge-mcp-workflows skill, which routes catalogue, governance, access and data quality questions through the server's playbooks. | Source V |
| Ataccama ONE | The Data Quality and Catalog MCP server, at your Ataccama ONE host followed by /mcp (from version 17.2.0), signed in with Ataccama ONE credentials. Ataccama Cloud only, enabled on request; self-managed deployments are not supported. | Not listed | Source S | |
| DataGalaxy | A read-only MCP server that searches objects, links, users and tags, returns ownership, definitions and lineage, and keeps DataGalaxy's roles and access controls. Its address is shown in each DataGalaxy instance, with a personal token. | Not listed | Source S | |
| Microsoft Purview (Data Map and Unified Catalog) | No Microsoft-published Purview MCP server was found; community servers exist. Not listed here until one is checked. | Not listed | K | |
| Apache Atlas | No Apache-published MCP server was found. | Not listed | K | |
| OpenMetadata (and Collate) | A built-in MCP server at /mcp on every OpenMetadata server since 1.8.0. | search_metadata | The AI SDK turns the MCP tools into LangChain and OpenAI function tools, and calls Collate's ready-made agents. | Source V |
| DataHub (and DataHub Cloud) | An MCP server over stdio or HTTP, with search, lineage and editing tools. | search, get_entities, get_lineage, get_lineage_paths_between, list_schema_fields, get_dataset_queries, update_description, add_tags, add_terms, add_owners, set_domains, add_structured_properties | Source V | |
| Unity Catalog (open source and Databricks) | Databricks registers MCP servers in Unity Catalog as governed objects; a Unity Catalog metadata MCP server published by the project was not found. | Not listed | K | |
| Google Cloud Dataplex Universal Catalog | MCP Toolbox for Databases has a prebuilt Dataplex tool set. | search_entries, lookup_entry, lookup_context, search_dq_scans, list_data_products | The toolbox can generate agent skills from its tool sets. | Source V |
| Snowflake (Horizon Catalog and Snowflake Open Catalog) | The official Snowflake MCP Server, managed by Snowflake and built on Cortex Agents; the open-source Snowflake-Labs/mcp server is deprecated and points to it. | Not listed | Cortex Analyst answers questions against semantic views; Cortex Search retrieves from unstructured data. | Source V |
| AWS Glue Data Catalog | The AWS Data Processing MCP server (awslabs.aws-dataprocessing-mcp-server), local over stdio with an AWS profile; changes need the --allow-write flag. | manage_aws_glue_databases, manage_aws_glue_tables, manage_aws_glue_partitions, manage_aws_glue_catalog, manage_aws_glue_connections, manage_aws_glue_crawlers | Source V | |
| Amazon SageMaker Catalog (Amazon DataZone) | No AWS-published MCP server for SageMaker Catalog was found. | Not listed | K | |
| Oracle Cloud Infrastructure Data Catalog | No Oracle-published MCP server for OCI Data Catalog was found. Oracle Analytics Cloud has an MCP server in preview; it covers analytics datasets | Not listed | K | |
| SAP Datasphere | Community MCP servers only, not published or supported by SAP. They read spaces and catalogue assets through Datasphere's Catalog and Consumption APIs. | Not listed | Source V | |
| Cloudera (SDX) | No Cloudera-published MCP server for its catalogue was found. | Not listed | K | |
| Teradata Vantage | An MCP server installed with uv or pip, connecting with a database URI; it reads schemas and column metadata and runs SQL. | Not listed | Source V | |
| Qlik Talend Cloud (Talend Data Catalog) | An MCP server in Qlik Talend Cloud, switched on per tenant by an administrator, giving AI assistants governed data sets and data products. | Not listed | Source S | |
| Precisely Data Integrity Suite | Reported to host an MCP server over its integration | Not listed | K | |
| Select Star | A hosted MCP server at mcp.production.selectstar.com/mcp, used with a Select Star API token. | Not listed | Source S | |
| Secoda | A hosted MCP server in each Secoda workspace, and an open-source server that uses a Secoda API token. | search_data_assets, retrieve_entity, entity_lineage, search_documentation, run_sql | Source V | |
| Actian Data Intelligence Platform (formerly Zeenea) | A managed MCP server connecting AI assistants such as Claude and Microsoft Copilot to the platform's metadata. | Not listed | Source S | |
| Apache Polaris (Iceberg REST catalog) | No Apache-published MCP server was found. | Not listed | K | |
| OpenLineage (and Marquez) | OpenLineage is a specification for lineage events; it has no MCP server of its own. | Not listed | K | |
| Amundsen | No Amundsen-published MCP server was found. | Not listed | K | |
| Egeria (open metadata) | No Egeria-published MCP server was found. | Not listed | K |
In Claude
Connectors in the Claude connector directory, and plugins and skills in the plugin directories and the vendors' own repositories, searched on 2026-11-07. A connector needs the organisation’s own tenant and sign-in; a plugin bundles skills that tell Claude how to use the vendor’s tools well. Nothing was installed or connected. The directory search covers only part of the directory, so a missing entry may still exist. None found for Alation Data Intelligence Platform, IBM watsonx.data intelligence, data.world, Apache Atlas, OpenMetadata (and Collate), Ataccama ONE, DataGalaxy, Amazon SageMaker Catalog (Amazon DataZone), Oracle Cloud Infrastructure Data Catalog, SAP Datasphere, Cloudera (SDX), Teradata Vantage, Precisely Data Integrity Suite, Select Star, Secoda, Actian Data Intelligence Platform (formerly Zeenea), Apache Polaris (Iceberg REST catalog), OpenLineage (and Marquez), Amundsen, Egeria (open metadata).
| Product | In Claude | Specification | What it is for | Source |
|---|---|---|---|---|
| Atlan | Atlan Plugin. Anthropic Directory, partner publisher atlanhq | MCP server atlan, http, https://mcp.atlan.com/mcp, OAuthSkills: atlan-search: Chooses between semantic search | Find, explain and govern Atlan assets from Claude Code; lineage, glossary terms, certification and data quality rules. | atlanhq/agent-toolkit (commit 17b7ae3b4852) |
| Collibra Platform | Collibra Connector. Claude connector directory | MCP server Collibra, http, https://<your instance>/rest/mcp, OAuth per userTools: discover_data_assets, discover_business_glossary, create_asset, edit_asset, create_assessment, edit_assessment, get_assessment, add_data_classification_match (The directory lists eight tools by name and 21 more.) | Search governed assets and the business glossary, follow lineage, and create or edit assets within Collibra's own permissions. | |
| Informatica Cloud Data Governance and Catalog (CDGC) | Informatica (informatica-for-claude-platform) Plugin. Knowledge Work plugins; enabled on this account | MCP server informatica-catalog-discovery, informatica-data-exploration, http, Set per IDMC tenant, Tenant credentialsSkills: catalog-and-data-discovery: Finds tables | Ground an agent in CDGC before it queries data, so it picks the trusted, compliant asset. | forcedotcom/informatica-claude-plugins (commit 554d02e0f7f2) |
| Informatica Cloud Data Governance and Catalog (CDGC) | Informatica Catalog Discovery Connector. Claude connector directory | Tools: search_assets, get_asset_details | Discover and explore assets in Cloud Data Governance and Catalog. | |
| Informatica Cloud Data Governance and Catalog (CDGC) | Informatica Data Exploration Connector. Claude connector directory | Tools: data_explore, master_data_explore, get_skill_execution_updates, request_upload_url | Explore dataset content and mastered (MDM) entities in IDMC through Informatica's discovery agent. | |
| OvalEdge | claude-ovaledge Plugin. In the vendor's own repository (not in the Claude directories searched) | MCP server OvalEdge, stdio or http, https://<your OvalEdge server>/mcp, API key or single sign-onSkills: ovaledge-mcp-workflows: Routes catalogue | Install from the repository's own plugin marketplace to search, explain and govern OvalEdge assets from Claude Code. | ovaledge/oe_mcp (commit 9587d7c18075) |
| Microsoft Purview (Data Map and Unified Catalog) | Microsoft Learn Connector. Claude connector directory; connected on this account | Tools: microsoft_docs_search, microsoft_docs_fetch, microsoft_code_sample_search | Look up Purview's own documentation, such as the Atlas typedef and business metadata APIs. It does not reach a Purview account. | |
| Microsoft Purview (Data Map and Unified Catalog) | microsoft-docs Plugin. Anthropic Directory, partner publisher microsoftdocs | Skills: microsoft-docs: Answers from official Microsoft documentation.; microsoft-code-reference: Finds API references and code samples. | The same documentation, packaged with skills. | microsoftdocs/mcp (commit f0a9478a954c) |
| DataHub (and DataHub Cloud) | DataHub Connector. Claude connector directory | Tools: add_owners, add_related_terms, add_structured_properties, add_tags, add_terms, create_glossary_term, compare_glossary_term_versions, accept_or_reject_proposals (The directory lists eight tools by name and 26 more.) | Search and enrich the catalog, manage glossary terms and structured properties, and review proposals. | |
| DataHub (and DataHub Cloud) | DataHub Cloud Plugin. Anthropic Directory, partner publisher datahub-project | MCP server datahub, http, https://mcp.datahub.com/mcp, OAuthSkills: datahub-search: Finds datasets; datahub-lineage: Traces what feeds and depends on an asset; datahub-quality: Creates and runs assertions; datahub-sql-workflow: Grounds text-to-SQL in catalog evidence before writing a query.; datahub-setup: Checks the MCP connection and authentication. | Catalog search, lineage, quality and grounded SQL on DataHub Cloud. | datahub-project/datahub-cloud-ai-plugins (commit 6d997e960d3e) |
| DataHub (and DataHub Cloud) | datahub-skills Plugin. Anthropic Directory, partner publisher datahub-project | Skills: datahub-connector-planning: Plans a new DataHub ingestion connector.; datahub-connector-pr-review: Reviews a connector against DataHub's golden standards.; datahub-enrich: Adds descriptions | Building and reviewing DataHub connectors, which is how a DCAT-AP-UK source would be ingested. | datahub-project/datahub-skills (commit a27f8e7bb514) |
| Unity Catalog (open source and Databricks) | databricks Plugin. Anthropic Directory, partner publisher databricks | Skills: databricks-unity-catalog: Grants and revokes access; databricks-data-discovery: Finds where data lives and answers questions about it through Genie. | Govern Unity Catalog objects and read their lineage from Claude Code. | databricks/databricks-agent-skills (commit e6c32c58fbfe) |
| Unity Catalog (open source and Databricks) | Databricks Genie Connector. Claude connector directory | Tools: genie_ask, genie_poll_response, genie_get_query_result, genie_cancel_response, view_ask | Ask questions of data that Unity Catalog governs, through a Genie space. | |
| Google Cloud Dataplex Universal Catalog | Google Cloud BigQuery Connector. Claude connector directory | Tools: list_dataset_ids, get_dataset_info, list_table_ids, get_table_info, execute_sql | Read BigQuery dataset and table metadata. It does not expose Dataplex catalog entries or aspects; for those use MCP Toolbox's Dataplex tools above. | |
| Snowflake (Horizon Catalog and Snowflake Open Catalog) | Snowflake Connector. Claude connector directory | Tools: Cortex Agents, Cortex Search | Ask questions of governed Snowflake data through Cortex Agents, and retrieve unstructured content with Cortex Search. | |
| Snowflake (Horizon Catalog and Snowflake Open Catalog) | snowflake-cortex-code Plugin. Anthropic Directory, partner publisher snowflake-labs | Skills: cortex-router: Sends Snowflake requests from Claude Code to Cortex Code to run.; cortex-run: Runs a Snowflake task through Cortex Code.; cortex-setup: Sets up the connection to Cortex Code. | Run Snowflake work from Claude Code, including object and catalogue management. | snowflake-labs/snowflake-ai-kit (commit 6a9165f38da7) |
| AWS Glue Data Catalog | aws-data-analytics Plugin. Anthropic Directory, partner publisher aws | MCP server aws-mcp, stdio, Runs with your AWS credentials, AWS profileSkills: exploring-data-catalog: Inventories and audits Glue Data Catalog assets; finding-data-lake-assets: Resolves references to data lake and lakehouse assets across the Glue Data Catalog and S3. | Explore and audit the Glue Data Catalog from Claude Code. | aws/agent-toolkit-for-aws (commit 0a218f8cc132) |
| Qlik Talend Cloud (Talend Data Catalog) | Qlik Connector. Claude connector directory | Tools: qlik_create_app, qlik_create_data_object, qlik_add_chart, qlik_add_filter, qlik_create_bookmark, qlik_create_automation, qlik_clear_selections, qlik_cancel_app_reload (The directory lists eight tools by name and 85 more.) | Work with Qlik Cloud apps, data objects and automations from Claude, grounded in governed Qlik data. |
Add-on kits
Definitions of every DCAT-AP-UK dataset property, in each product's own extension format, named as the weaver writes them.
- Business metadata typedef (Apache Atlas, Microsoft Purview, Atlan)
- Import API template and attribute types (Collibra)
- Dataset properties template (data.world)
- Custom properties for the table entity (OpenMetadata)
- Structured properties (DataHub)
- Table properties (Unity Catalog)
- Aspect type (Dataplex Universal Catalog)
- Table properties (Apache Iceberg REST catalogs: Snowflake Open Catalog, Apache Polaris)
- Table parameters (AWS Glue Data Catalog)
- Metadata form type in Smithy (Amazon SageMaker Catalog, DataZone)
- Custom dataset facet schema (OpenLineage)
- Custom properties (OCI Data Catalog)
The reports
| Report | By | Published | Leaders, with the announcement that says so | Note |
|---|---|---|---|---|
| Magic Quadrant for Metadata Management Solutions | Gartner | 2025-11-19 | Informatica Cloud Data Governance and Catalog (CDGC) S, IBM watsonx.data intelligence S, Alation Data Intelligence Platform S, Ab Initio Software K, Atlan S | The first Gartner evaluation of this market since 2020, according to Informatica's announcement. Vendors named in the published abstract include Ab Initio Software, Alation, Alex Solutions and Ataccama. |
| Magic Quadrant for Data and Analytics Governance Platforms (first edition) | Gartner | 2025-01-07 | Informatica Cloud Data Governance and Catalog (CDGC) S, IBM watsonx.data intelligence S, Collibra Platform S | The first Magic Quadrant for this market. The abstract names Ab Initio Software, Alation, Alex Solutions, Anjana Data, Ataccama, Atlan, Collibra, DataGalaxy, data.world, erwin by Quest, Global Data Excellence, IBM, Informatica and OvalEdge; press coverage also names Solidatus. A second edition followed on 7 January 2026. |
| Magic Quadrant for Data and Analytics Governance Platforms (second edition) | Gartner | 2026-01-07 | IBM watsonx.data intelligence S, Atlan S | Known from vendor announcements; other vendors' positions in this edition are not yet checked. |
| The Forrester Wave: Data Governance Solutions, Q3 2025 | Forrester | 2025-07-09 | Alation Data Intelligence Platform S, Atlan S, Collibra Platform S, Informatica Cloud Data Governance and Catalog (CDGC) S | Thirteen providers evaluated. |
| The Forrester Wave: Enterprise Data Catalogs, Q3 2024 | Forrester | 2024-08-08 | Atlan S, Collibra Platform S, Informatica Cloud Data Governance and Catalog (CDGC) S | Twelve providers evaluated against 24 criteria. |
Not covered yet
Named in a report above, with no public metadata model or MCP server found on 2026-11-07. The weekly run looks again.
- Ab Initio Software (Gartner metadata management 2025, Gartner D&A governance 2025)
- Alex Solutions (Gartner metadata management 2025, Gartner D&A governance 2025)
- Anjana Data (Gartner D&A governance 2025)
- erwin by Quest (Gartner D&A governance 2025)
- Global Data Excellence (Gartner D&A governance 2025)
- Solidatus (Gartner D&A governance 2025) Has an AI lineage assistant; no MCP server or public model found. Source.