MetaKarta v12 Unifies Enterprise Metadata Management for AI, BI, and Databases

The expanded metadata management platform combines data lineage, data catalog, data governance and semantics on a shared foundation, with compiled context for AI agents and native artifacts for databases and BI.

MOUNTAIN VIEW, Calif., Sept. 22, 2026 /PRNewswire/ -- Meta Integration Technology, Inc. (MITI) today announced general availability of MetaKarta v12, including Semantic Hub that compiles governed business definitions into the databases, BI tools, and AI agents that depend on them. A definition authored once in MetaKarta lands as a native artifact inside Snowflake, Databricks, Power BI, and Tableau, and reaches AI agents as compiled context drawn from the same repository, so a finance dashboard and an AI agent answering the same question return the same number. MetaKarta v12 is built to produce three outcomes from one system: trusted BI, reliable AI, and defensible governance.

MITI has built OEM metadata infrastructure for nearly 30 years, and its technology ships today inside products from Microsoft Purview, Informatica from Salesforce, IBM, Oracle, and Qlik Talend. MetaKarta v12 brings that depth and breadth of metadata engineering to enterprises directly, with 400+ native connectors and parser-based, column-level lineage computed from actual code.

Now in v12, organizations own the semantic artifacts compiled and deployed in their own databases and BI tools. Definitions stay portable through a migration, for example from SAP BusinessObjects to Power BI, or Oracle to Snowflake or Databricks. The compiled definitions are already in place before any query runs, so no MetaKarta proprietary execution engine sits in the query path. Those same definitions are delivered as compiled context to AI agents.

MetaKarta v12 also reflects a shift MITI calls MetadataOps: metadata management moving from periodic projects to a continuous enterprise operation. As data, applications, definitions, and AI systems change, the metadata connecting them has to change with them. MetaKarta provides the shared foundation for keeping that work continuous across data lineage, data catalog, data governance, and semantics.

"For most of our history we built OEM components, and that worked while the problem lived inside one platform," said Christian Bremeau, Founder and CEO of Meta Integration Technology. "AI moved it. Most enterprises now point agents at metadata fragmented across a lineage tool, a catalog, a governance tool, and multiple semantic layers, so the agent picks whichever definition sits nearest the data and acts on it at machine speed, with no analyst catching the discrepancy."

"That's why we think metadata management now has to operate as MetadataOps, a continuous discipline rather than a series of disconnected projects," Bremeau continued. "We built v12 so the business meaning underneath an agent is trustworthy before it ever acts. Definitions have to be portable, vendor-neutral, and owned by the enterprise that depends on them. When reports, dashboards, and AI agents all start from unified metadata, people and agents get the same answer wherever the question comes from."

What's inside MetaKarta v12
MetaKarta v12 expands the metadata management platform to include Semantic Hub, platform-wide AI, and a fully redesigned user experience.

• Semantic Hub: Reverse-engineer existing semantic assets from Power BI, Tableau, Looker, MicroStrategy, SAP BusinessObjects, IBM Cognos, and Oracle OBIEE, govern them centrally, then compile them into native artifacts: Snowflake Semantic Views, Databricks Metric Views, Oracle Analytic Views, Power BI semantic models, Tableau logical models, and LookML.
• Semantic Hub Language (SHL): A YAML-based language for defining semantic models.
• MetaKarta MCP for Agentic AI: Metadata Management Tools provide access to governed metadata, while Semantic Hub Tools deliver compiled context and deterministic SQL to AI agents from the shared metadata repository. Per-user access tokens mean an agent sees exactly what the person behind it is authorized to see.
• Context Sandbox: Validate AI grounding context before deployment, tracing an answer from source to output.
• Ontology Modeling: Define business entities, relationships, and hierarchies in MetaKarta Business Glossary as a governed ontology, bound to the semantic models and physical columns that implement them.
• Ask MetaKarta: Natural-language access to governed metadata for business and technical users.

What MetaKarta changes

• Trusted BI. When "revenue" is coded separately in Power BI and in Tableau, finance and sales reconcile the difference before anyone can act on the number. One governed definition compiles into both.
• Reliable AI. An agent querying raw schema at run-time infers what a metric means from column names, and infers it differently every time. Compiled context supplies the canonical definition and the provenance behind it, alongside structured AI consumption inside Snowflake, Databricks, and Oracle.
• Defensible Governance. Auditors ask what a field meant on a specific date, how it was calculated, and who approved the change. Version-controlled definitions, lineage, and audit-ready metadata turn that request into a query.

Availability

MetaKarta v12 is generally available today. Full details are available at www.metakarta.com.

The MetadataOps Kit is available at www.metakarta.com/resources/metadataops, covering the practice, operating lifecycle, metadata quality scoring model, a paste-ready job description and skills matrix for the MetadataOps Engineer role, and more.

About Meta Integration Technology

Meta Integration Technology, Inc. (MITI) is a pioneer in metadata integration, semantic modeling, and data lineage, built on nearly 30 years of experience serving global enterprises and technology providers. MITI's technology is embedded in products from Oracle, Microsoft Purview, Informatica from Salesforce, IBM, and Qlik Talend, and it powers MetaKarta, MITI's enterprise platform for metadata-driven data management. MetaKarta delivers data lineage, data catalog, data governance, and semantic compilation through a single integrated system, connecting the full enterprise data estate across 400+ native connectors, from legacy to modern and on-premises to cloud.

SOURCE Meta Integration Technology