FEATURE
Snowflake Horizon Catalog
Govern and find data and AI assets across Snowflake and external catalogs. Enterprise context, lineage, data quality monitoring and AI guardrails mean every answer is traceable and trusted.
Connect your entire data estate
One catalog for data inside and outside Snowflake. Open table formats. Any compute engine.
Build your AI context layer, automatically
Collect, enrich and activate context across BI and data so humans and AI operate on the same trusted semantics.
Deploy governed, trustworthy AI
AI guardrails detect and block sensitive data before it reaches users. End-to-end lineage traces AI-generated answers back to the source.
Collaborate withoutmoving your data
Whether working together on data with a publisher or advertiser, or sharing between teams or across departments, data clean rooms allow you to derive insights without worrying about the risks data movement can bring.
Interoperable lakehouse catalog
Built on Apache Polaris, Horizon Catalog uses Iceberg REST and Scan APIs to enable any engine to query governed data in place with bidirectional read and write across Snowflake and external catalogs.
Search and discovery
Search your entire catalog in natural language via Snowflake CoCo, spanning tables, dashboards, notebooks, agents, and external BI metadata in one place. Share AI-ready data products across regions and clouds in a few clicks.
Context for AI and BI
Surface and enrich context with external metadata connectors, automated semantic view creation, popularity-based asset ranking, and unified context search.
AI governance and security
Enforce AI guardrails, manage agent policies, quality, and MCP from a centralized dashboard, and use natural language to ensure every AI agent operates within your security boundaries.
Data observability
Ensure data trust with continuous quality monitoring and column-level lineage that gives visibility from source to BI, tracing AI-generated answers to origin for root cause analysis.
Sensitive data protection
Automatically classify and tag sensitive data, then enforce protection at scale with continuous monitoring, row-level filtering, and column masking.
BENEFITS
Horizon Catalog: Built for trusted answers
Universal Agentic Catalog
Ask Snowflake CoCo in plain language. Get trusted responses based on your data.
Convert intent into active Horizon Catalog policies for masking, access controls, data quality and more.*
Classify, tag and add documentation and even create semantic models with a simple prompt.
Query data across Snowflake, external data lakes and external relational databases without switching tools.
Connect to any Iceberg catalog to discover and act on fresh data — reads and writes.
A connected data estate
A truly open and interoperable catalog
Benefit from community contributions by choosing the catalog implementing open APIs from Iceberg and Apache Polaris™.
Work off a single governed copy of your data, on any catalog, for read and write operations from Snowflake.
Integrate data sources like external databases, BI and data pipeline systems quickly with out-of-the-box connectors, including PostgreSQL, Tableau, Microsoft Power BI, dbt and more.
A single catalog UI surfaces Snowflake objects, external tables, dashboards and AI assets in one place.
AI context layer, built automatically
Collect, enrich and activate context so humans and AI operate on the same trusted semantics
Gather rich context, such as query logs, popularity, joins and relationships from Tableau, Power BI and other BI sources into one searchable catalog where every engine sees the same definitions and lineage.
Track column-level lineage across Snowflake and external databases, rank assets by popularity and auto-generate documentation from metadata.
Semantic View Autopilot creates business-ready definitions from your existing BI sources; Semantic Studio adds git versioning, and CoCo edits in one workspace.
Open Semantic Interchange keeps business logic consistent across any tool or catalog.
Deploy governed, trustworthy AI
Every answer is backed by secure, policy-enforced data
Agents operate under the same RBAC policies as humans. An agent dashboard shows all active agents, MCP connections and policy status.
AI Guardrails detect, redact and block PII from agent outputs before they reach users or tools.
Data quality monitoring and lineage ensure every query hits fresh data.
Masking and row-access policies defined in Snowflake are enforced for data queried through any Iceberg REST Catalog-compatible external engine.
Sensitive data monitoring discovers and classifies data, applying tags and policies that keep PII safe.
Advertising, Media and Entertainment
Merkle Improves Customer Experiences While Providing Data Governance and Security
Merkle, a dentsu company, consolidates sensitive data and collaborates with clients in Snowflake, resulting in a more efficient, trusted data environment that expedites data access and reduces risk.
- 64% Faster data development cycle
- 20% Estimated cost savings
Your Governance Resources
Resources, partner integrations and community expertise to help you govern data at scale.
See Horizon Catalog in Action
See live demos of Horizon Catalog bi-directional interoperability, governance for AI and semantic views.
Snowflake Community
Meet and learn from a global network of data governance leaders in Snowflake’s community forum and Snowflake User Groups.
Snowflake Documentation
Explore Snowflake documentation for features, tutorials and a detailed reference of commands and operations.
Resources
