Select Star.

What Select Star is

Select Star is an automated data catalog and lineage platform — the company’s own framing is “metadata context platform” — founded in San Francisco in 2020 by Shinji Kim. Its core mechanic: connect read-only to your warehouse and BI tools, parse the SQL query logs and dbt manifests, and generate column-level lineage, popularity metrics, entity-relationship diagrams, and suggested documentation automatically. Around that core sit Ask AI (natural-language search over the catalog), an MCP server that exposes metadata and lineage to LLM agents, automated semantic model generation, and a Data Access Control module with automated PII tagging, tag propagation, and RBAC/ABAC policies. On November 24, 2025, Snowflake announced a definitive agreement to acquire the team and platform technology, with the stated goal of expanding Horizon Catalog for agentic AI experiences like Snowflake Intelligence and Cortex Code.

Where it fits

Select Star cross-shops with atlan, secoda, alation, and collibra among proprietary catalogs, and with datahub, openmetadata, and amundsen on the open-source side — its own site maintains comparison pages against most of these. Its wedge has always been automation-first lineage: where enterprise suites lead with stewardship workflows and OSS catalogs lead with portability, Select Star leads with column-level lineage extracted from query logs with minimal setup, proven at unusual scale (Block used it to automate lineage across an exabyte-scale warehouse). It consumes OpenLineage events, so Airflow and pipeline lineage land in the same graph as warehouse and BI lineage. What it is not is a data-quality engine: there are no monitors, tests, or anomaly detection, which is why the quality-testing cluster scores zero — pair it with a dedicated tool if you need runtime checks.

On the Snowflake acquisition

The announcement is unambiguous about direction: Select Star’s lineage and discovery capabilities are headed into Snowflake’s Horizon Catalog. That cuts both ways. For Snowflake-centric organisations it is arguably good news — best-in-class automated lineage becoming a native platform capability. For everyone else it introduces real uncertainty: the public pricing page has already been removed, and the long-term standalone product, its cross-platform integrations, and its commercial terms are all open questions until the transition settles. Kim’s announcement says the multi-source integrations will be brought into Horizon Catalog, but a Snowflake-owned catalog’s incentives on BigQuery or Databricks coverage are not the same as an independent vendor’s.

How to evaluate it

Scope any trial around the automation claim, because that is the whole pitch: connect one warehouse, one dbt project, and one BI tool, then check whether column-level lineage materialises accurately across all three without manual stitching, and whether the auto-generated documentation and popularity signals actually help analysts find the right table. Then ask the acquisition questions directly — standalone availability for new customers, support commitments for non-Snowflake sources, and what happens to your contract when the technology folds into Horizon Catalog. If you are already all-in on Snowflake, the safest read may be to evaluate Horizon Catalog’s roadmap rather than Select Star as a standalone purchase.