How to Find, Trace, and Govern Your Metadata in MetaKarta
The User Guide is for data stewards, analysts, and governance leads who work in MetaKarta every day. It covers searching and asking questions of your metadata, reading object pages and lineage, importing models, building glossaries with review workflows, cataloging data, and administering users, roles, and labels. When you're done, you can trace a column from source to report, show who approved a definition and when, and answer "what does this field mean?" with the governed definition.
Note
Some of the features detailed in this document may not apply and/or be available for the particular edition/version you are using.
Get started
Frequently asked questions
How do I search for a table or column?
Type in the search box in the upper right of any page and press Enter. Certified objects list first.
Search uses Apache Lucene syntax, so you can match an exact phrase, require every word with +, exclude a word, or use a trailing wildcard. An administrator can also assign personas that scope results to the models you care about.
Does MetaKarta support natural-language questions about metadata?
Yes. Ask MetaKarta answers natural-language questions about lineage, impact analysis, catalog entries, and definitions, using the default AI provider your administrator sets up.
You can ask where a field lives in Snowflake, which Power BI reports source a table, or what breaks if you rename something. Answers resolve against the governed metamodel and link straight to the objects.
See Ask MetaKarta.
How do I trace column-level lineage and impact?
Open lineage from the object and trace in reverse to see its sources, or forward to see its impact. MetaKarta computes lineage at the column level, shows data flow as solid lines and semantic flow as dashed lines, and offers tree and diagram views. You'll need a configuration that holds every model in the path, with connections resolved.
How do I import metadata into MetaKarta?
You create a model, pick the import bridge that matches the source technology, fill in its connection parameters, and run a full or incremental harvest. MetaKarta can create a new version only when the source changed. If a bridge needs Windows or a separate network, a remote harvesting agent does the import.
How do I set up review and approval for glossary terms?
Workflow is off by default, so edits appear immediately. Turn it on and terms move through Draft, Under Review, Approved, and Published, with editor, reviewer, approver, and publisher roles.
MetaKarta then keeps a Development version for editing and a Published version for everyone else. Labels, attachments, and comments stay editable outside workflow.
How do I tag and protect personally identifiable information (PII)?
Create a compound PII data class from the predefined classes (last name, postal code, Social Security number, and others) and set its sensitivity to Confidential. During harvesting, MetaKarta tags matching fields and hides their profiling data from users without the right permissions. Link the class to a glossary term to connect the business definition to every tagged field.
What does the User Guide cover for Semantic Hub?
The Semantic Hub section walks through importing semantic models from databases and BI tools, modeling and reconciling them in Semantic Hub Language, compiling them into native database and BI artifacts, and binding them to a glossary-based ontology. It also covers serving compiled context to AI agents through the Semantic Hub Tools in MetaKarta MCP, and testing that context in Context Sandbox.
Related documentation
Return to the MetaKarta documentation home To run these features as a guided demo, see the Demo Guide; to query metadata with MQL or the REST API, see the Developer Guide.