PLATFORM PROFILE, read 13 Nov 2026
DataHub for AI data governance: rank 3 of 16, at 61.0 of 100
See the full indexOpen the exposure map
- Category
- Open source data catalog and metadata platform with a managed Cloud edition
- Status
- Exists under this name on 13 Nov 2026 (datahub.com and docs.datahub.com live; open source DataHub by DataHub Inc, formerly Acryl Data; the company page and Series B press release say 'DataHub, by Acryl Data'; security page still says Acryl; GitHub repo pushed 13 Nov 2026).
- Exposure trail
- Warehouse column: partial; Transformation: partial; BI field: partial; Report: partial; Viewer: partial; Agent: documented
- Warehouses named
- Snowflake, BigQuery, Databricks, Redshift
- BI tools named
- Tableau, Power BI, Looker, Preset
- Security tools named
- none found on 13 Nov 2026
- Rank and total
- 3 of 16, 61.0 of 100
1 of 6 stages documented, 5 partial
DataHub on the eight criteria
| Criterion | Weight | Score | Why |
|---|---|---|---|
| Follows sensitive data into BI | 25 | 8 | Column lineage from Snowflake through dbt into Looker, Power BI and Tableau is documented, and dbt Labs independently describes column lineage to downstream dashboards; Tableau needs a high role. |
| Controls what an agent is told | 20 | 7 | Per-user OAuth tokens keep permissions and audit identity per user, confirmed in the public README; scoped MCP servers are per use case and need an add-on. |
| Shows who actually saw it | 15 | 4 | Looker historical usage and per-user audit attribution exist, but viewer lists for Power BI or Tableau reports were none found on 2026-11-13. |
| Carries classification downstream | 15 | 6 | Term and tag propagation along column lineage with a BigID connector (alpha), but datasets only, Public Beta, and the built-in PII classifier was removed. |
| Acts on it | 10 | 4 | Metadata tests apply labels automatically and incidents can block pipelines, but access workflows are Private Beta and no masking is shown. |
| Coverage of the stack | 5 | 6 | Generally available Snowflake, BigQuery, Databricks, dbt, Looker, Tableau and Power BI connectors; security coverage is only an alpha BigID connector. |
| Evidence beyond the vendor | 5 | 9 | Airflow and dbt Labs list it, and public repository, PyPI and Docker Hub records show active, widely used code. |
| Price transparency | 5 | 1 | Cloud is sold through sales with no price list found on 2026-11-13; only the free Core edition is stated. |
Where DataHub is strongest
Vendor-stated: Column-level lineage tracking, glossary term propagation to downstream lineage columns, and per-user OAuth logins for MCP.
Where DataHub falls short
The built-in PII classifier was removed from ingestion, asset-level term propagation skips dashboards and charts, and Cloud pricing is not public.
DataHub along the exposure trail
| Stage | Reading | Evidence |
|---|---|---|
| Warehouse column | partial | DataHub docs say the built-in PII classifier was removed from ingestion; no classifier is registered by default. Source, read 13 Nov 2026. |
| Transformation | partial | DataHub docs say column-level lineage tracks each specific column, shown inside table-level lineage. Source, read 13 Nov 2026. |
| BI field | partial | The Looker source docs list column-level lineage as enabled by default. Source, read 13 Nov 2026. |
| Report | partial | DataHub docs limit asset-level glossary term propagation to datasets; charts, dashboards, pipelines and tasks are excluded. Source, read 13 Nov 2026. |
| Viewer | partial | The Looker source docs say it can pull historical dashboard usage from Looker system activity. Source, read 13 Nov 2026. |
| Agent | documented | DataHub docs say Cloud admins can create scoped MCP servers with their own URL, hidden tools, custom instructions and a saved View; this needs the Context add-on. Source, read 13 Nov 2026. |
DataHub in each situation
| Situation | Rank | Total | Leader |
|---|---|---|---|
| A regulated field | 3 | 63.0 | Atlan 68.5 |
| Agents on analytics | 3 | 59.5 | Euno 66.0 |
| A Microsoft estate | 3 | 61.0 | Atlan 68.5 |
| A mixed stack | 3 | 61.5 | Euno 68.5 |
| A lean team | 5 | 53.0 | Euno 62.5 |
The evidence on DataHub, fact by fact
- datahub-01 (Profile, DataHub's own claim): DataHub docs say Metadata Tests apply their actions (tags, terms, owners, domains) continuously as the set of matching assets changes, with no person marking each asset. Source, read 13 Nov 2026.
- datahub-02 (Profile, DataHub's own claim): DataHub docs say a Metadata Test action can add or remove classification labels as assets move in and out of compliance with a rule. Source, read 13 Nov 2026.
- datahub-03 (Profile, DataHub's own claim): DataHub docs say Glossary Term Propagation copies classification labels to downstream columns and assets along lineage and siblings. Source, read 13 Nov 2026.
- datahub-04 (Profile, DataHub's own claim): The glossary term propagation page labels the feature as Public Beta in DataHub Cloud, available through a Cloud representative. Source, read 13 Nov 2026.
- datahub-05 (Carries classification downstream, DataHub's own claim): DataHub docs limit asset-level glossary term propagation to datasets; charts, dashboards, pipelines and tasks are excluded. Source, read 13 Nov 2026.
- datahub-06 (Carries classification downstream, DataHub's own claim): DataHub docs say column-level term propagation covers all downstream lineage columns and sibling columns, and a chosen subset of terms (for example only sensitive ones) can be selected. Source, read 13 Nov 2026.
- datahub-07 (Carries classification downstream, DataHub's own claim): DataHub docs describe a Cloud-only automation that syncs DataHub terms, tags and descriptions back to Snowflake tables and columns (Public Beta). Source, read 13 Nov 2026.
- datahub-08 (Carries classification downstream, DataHub's own claim): The DataHub Actions framework includes a Tag Sync (Tag Propagation) action that propagates tags from an asset to downstream datasets, filtered by tag prefix, but only to datasets and not to dashboards or charts. Source, read 13 Nov 2026.
- datahub-09 (Profile, DataHub's own claim): DataHub docs say documentation propagation sends column descriptions to downstream and sibling columns by column-level lineage; it is on by default in open source and Public Beta in Cloud. Source, read 13 Nov 2026.
- datahub-10 (Profile, DataHub's own claim): DataHub docs say Metadata Test conditions can read technical signals such as usage and size that DataHub extracts, varying by integration. Source, read 13 Nov 2026.
- datahub-11 (Profile, DataHub's own claim): DataHub docs give example Metadata Test rules that require an owner, a description or a classification term on heavily queried assets. Source, read 13 Nov 2026.
- datahub-12 (Profile, DataHub's own claim): DataHub docs say Metadata Tests are re-evaluated when an asset changes and at least once every day. Source, read 13 Nov 2026.
- datahub-13 (Profile, DataHub's own claim): DataHub Cloud assertions can run a SQL check on a schedule against Snowflake, Redshift, BigQuery and Databricks, or evaluate ingested metadata (freshness, volume, schema). Source, read 13 Nov 2026.
- datahub-14 (Profile, DataHub's own claim): DataHub docs say Custom SQL and Column Value assertions need a live warehouse connection, so they do not work from ingested metadata alone. Source, read 13 Nov 2026.
- datahub-15 (Acts on it, DataHub's own claim): DataHub docs say incidents can be used to block data pipelines whose inputs have active issues. Source, read 13 Nov 2026.
- datahub-16 (Profile, DataHub's own claim): DataHub docs say its MCP Server gives AI agents direct access to DataHub metadata. Source, read 13 Nov 2026.
- datahub-17 (Profile, DataHub's own claim): DataHub docs say the MCP server returns usage stats, ownership, documentation, tags, glossary terms and quality signals for an asset. Source, read 13 Nov 2026.
- datahub-18 (Controls what an agent is told, DataHub's own claim): DataHub docs say DataHub Cloud MCP uses OAuth2 so each user connects with a personal login, with tokens scoped to that user. Source, read 13 Nov 2026.
- datahub-19 (Controls what an agent is told, DataHub's own claim): DataHub docs say a shared self-hosted MCP deployment keeps permissions, search results and audit attribution per user. Source, read 13 Nov 2026.
- datahub-20 (Controls what an agent is told, DataHub's own claim): DataHub docs say Cloud admins can create scoped MCP servers with their own URL, hidden tools, custom instructions and a saved View; this needs the Context add-on. Source, read 13 Nov 2026.
- datahub-21 (Acts on it, DataHub's own claim): DataHub docs say MCP mutation tools (add tags and terms, owners, description) are off unless TOOLS_IS_MUTATION_ENABLED is set. Source, read 13 Nov 2026.
- datahub-22 (Controls what an agent is told, observed): The public mcp-server-datahub README says the HTTP server has no server-wide token and each client sends its own DataHub token so permissions and audit identity stay per user. Source, read 13 Nov 2026.
- datahub-23 (Coverage of the stack, DataHub's own claim): DataHub says its integrations page covers 150+ data and AI systems, including 23 BI tools. Source, read 13 Nov 2026.
- datahub-24 (Follows sensitive data into BI, DataHub's own claim): DataHub docs show data flowing from Snowflake tables through dbt into Looker views and explores in its lineage view. Source, read 13 Nov 2026.
- datahub-25 (Follows sensitive data into BI, DataHub's own claim): DataHub docs say column-level lineage tracks each specific column, shown inside table-level lineage. Source, read 13 Nov 2026.
- datahub-26 (Follows sensitive data into BI, DataHub's own claim): The Looker source docs list column-level lineage as enabled by default. Source, read 13 Nov 2026.
- datahub-27 (Follows sensitive data into BI, DataHub's own claim): The Power BI source docs list column-level lineage as enabled by default. Source, read 13 Nov 2026.
- datahub-28 (Follows sensitive data into BI, DataHub's own claim): The Tableau source docs say a site role below Site Administrator Explorer limits data source field metadata and so column-level lineage. Source, read 13 Nov 2026.
- datahub-29 (Shows who actually saw it, DataHub's own claim): The Looker source docs say it can pull historical dashboard usage from Looker system activity. Source, read 13 Nov 2026.
- datahub-30 (Shows who actually saw it, none-found): None found on 2026-11-13: per-agent viewer or per-report viewer lists for Power BI or Tableau, in the DataHub docs for BI sources and usage. Searched: https://docs.datahub.com/docs/features/dataset-usage-and-query-history, read 13 Nov 2026.
- datahub-31 (Carries classification downstream, DataHub's own claim): DataHub docs say the built-in PII classifier was removed from ingestion; no classifier is registered by default. Source, read 13 Nov 2026.
- datahub-32 (Profile, DataHub's own claim): DataHub says on its governance page that AI-powered PII classification is part of DataHub Cloud workflows. Source, read 13 Nov 2026.
- datahub-33 (Profile, DataHub's own claim): DataHub lists GA connectors for Snowflake, BigQuery, Databricks, dbt, Looker, Tableau and Power BI on its integrations page. Source, read 13 Nov 2026.
- datahub-34 (Carries classification downstream, DataHub's own claim): DataHub lists a BigID connector (labeled Alpha) that enriches DataHub datasets with BigID classification: glossary terms, tags and risk scores. Source, read 13 Nov 2026.
- datahub-35 (Profile, DataHub's own claim): DataHub says its governance workflows track certification status in real time and flag non-compliant assets. Source, read 13 Nov 2026.
- datahub-36 (Acts on it, DataHub's own claim): DataHub docs list Access Workflows (approval routing for data access requests) as Private Beta. Source, read 13 Nov 2026.
- datahub-37 (Price transparency, DataHub's own claim): DataHub says DataHub Cloud pricing is scoped to use case and sold through sales; Core is free. Source, read 13 Nov 2026.
- datahub-38 (Price transparency, none-found): None found on 2026-11-13: a public price list or pricing page for DataHub Cloud, at datahub.com/pricing (returns Page Not Found) and in the Core vs Cloud page. Searched: https://datahub.com/pricing, read 13 Nov 2026.
- datahub-39 (Evidence beyond the vendor, observed): The public datahub-project/datahub repository is Apache-2.0 licensed with about 12,981 stars, last pushed 2026-11-13, latest tag v1.9.0rc5 (GitHub API). Source, read 13 Nov 2026.
- datahub-40 (Evidence beyond the vendor, observed): The acryldata/mcp-server-datahub repository is Apache-2.0, last pushed 2026-11-09, latest release v0.8.0 on 2026-10-21 (GitHub API). Source, read 13 Nov 2026.
- datahub-41 (Evidence beyond the vendor, observed): PyPI lists acryl-datahub version 1.7.0.20 uploaded 2026-11-12, source linked to datahub-project/datahub. Source, read 13 Nov 2026.
- datahub-42 (Profile, DataHub's own claim): DataHub announced a $35 million Series B led by Bessemer Venture Partners on May 21, 2025, as "DataHub, by Acryl Data". Source, read 13 Nov 2026.
- datahub-43 (Profile, DataHub's own claim): DataHub says it has 3000+ organizations, 3M+ monthly PyPI downloads and 16,000+ community members, and is headquartered in Palo Alto. Source, read 13 Nov 2026.
- datahub-44 (Profile, DataHub's own claim): DataHub says Metrics & Semantic Models, in Beta, is a catalog of metric definitions with owner, description and calculation, and with synonyms and instructions meant for agents. Source, read 13 Nov 2026.
- datahub-45 (Profile, DataHub's own claim): DataHub docs say the MCP server includes a tool to set lifecycle stage (proposed, approved, deprecated) on assets and terms, which a person or agent applies. Source, read 13 Nov 2026.
- datahub-46 (Evidence beyond the vendor, third-party): Apache Airflow lists DataHub on its ecosystem page as a metadata platform that can collect lineage and other metadata from Airflow. Source, read 13 Nov 2026.
- datahub-47 (Evidence beyond the vendor, third-party): dbt Labs lists "Acryl Data" (DataHub Cloud's former company name) on its integrations page for viewing dbt models, docs, test results and column-level lineage. Source, read 13 Nov 2026.
- datahub-48 (Evidence beyond the vendor, observed): Docker Hub shows the acryldata/datahub-gms image as active with more than 5.9 million pulls and an update on the read date. Source, read 13 Nov 2026.
- datahub-49 (Evidence beyond the vendor, observed): Artifact Hub lists a public DataHub Helm chart under the Data processing category. Source, read 13 Nov 2026.
Compare DataHub
Sources
- https://docs.datahub.com/docs/tests/metadata-tests, read 13 Nov 2026
- https://docs.datahub.com/docs/automations/glossary-term-propagation, read 13 Nov 2026
- https://docs.datahub.com/docs/automations/snowflake-metadata-sync, read 13 Nov 2026
- https://docs.datahub.com/docs/datahub-actions/src/datahub_actions/plugin/action/tag, read 13 Nov 2026
- https://docs.datahub.com/docs/automations/docs-propagation, read 13 Nov 2026
- https://docs.datahub.com/docs/managed-datahub/observe/assertions, read 13 Nov 2026
- https://docs.datahub.com/docs/incidents/incidents, read 13 Nov 2026
- https://docs.datahub.com/docs/features/feature-guides/mcp, read 13 Nov 2026
- https://docs.datahub.com/docs/features/feature-guides/scoped-mcp-servers, read 13 Nov 2026
- https://raw.githubusercontent.com/acryldata/mcp-server-datahub/main/README.md, read 13 Nov 2026
- https://docs.datahub.com/integrations, read 13 Nov 2026
- https://docs.datahub.com/docs/features/feature-guides/lineage, read 13 Nov 2026
- https://docs.datahub.com/docs/generated/ingestion/sources/looker, read 13 Nov 2026
- https://docs.datahub.com/docs/generated/ingestion/sources/powerbi, read 13 Nov 2026
- https://docs.datahub.com/docs/generated/ingestion/sources/tableau, read 13 Nov 2026
- Searched: https://docs.datahub.com/docs/features/dataset-usage-and-query-history, read 13 Nov 2026
- https://docs.datahub.com/docs/metadata-ingestion/docs/dev_guides/classification, read 13 Nov 2026
- https://datahub.com/products/data-governance/, read 13 Nov 2026
- https://docs.datahub.com/docs/managed-datahub/workflows/access-workflows, read 13 Nov 2026
- https://datahub.com/products/cloud-vs-core/, read 13 Nov 2026
- Searched: https://datahub.com/pricing, read 13 Nov 2026
- https://api.github.com/repos/datahub-project/datahub, read 13 Nov 2026
- https://api.github.com/repos/acryldata/mcp-server-datahub, read 13 Nov 2026
- https://pypi.org/pypi/acryl-datahub/json, read 13 Nov 2026
- https://datahub.com/news/series-b-announcement/, read 13 Nov 2026
- https://datahub.com/company/, read 13 Nov 2026
- https://docs.datahub.com/docs/features/feature-guides/metrics-and-semantic-models, read 13 Nov 2026
- https://airflow.apache.org/ecosystem/, read 13 Nov 2026
- https://www.getdbt.com/product/integrations, read 13 Nov 2026
- https://hub.docker.com/v2/repositories/acryldata/datahub-gms, read 13 Nov 2026
- https://artifacthub.io/api/v1/packages/helm/datahub/datahub, read 13 Nov 2026
Last reviewed 13 Nov 2026.