Governed Semantic Search
Find your "churn tables" even if none of them are called that — and with an audit trail.
Framework / standard: ISO/IEC 38507 (traced AI usage)
Discoverability should not depend on perfect naming. Linedat's semantic search finds your "churn tables" even if none of them are called that, and it does so in a governed way.
By meaning, not by name
Using embeddings of the asset's description and name, semantic search returns results by proximity of meaning, not by exact keyword match. It reduces the time your team wastes looking for data.
Governed and logged
Every semantic query is logged in the AI usage ledger, so discoverability is not a black box: there is a trail of what was queried, when and by whom.
The limits (what we do not claim)
Embeddings are generated by an external provider (with a cost per query) and stored and queried with pgvector inside PostgreSQL — there is no dedicated external search engine or local embedding implementation. The base keyword search uses substring matching (ILIKE), not a native full-text index.
How Linedat helps
Linedat gives you discoverability by meaning without giving up governance: you find the data even if you cannot remember its name, and there is an audit trail.
Related capabilities
When AI writes a description, the system knows AI wrote it — and does not publish it until a human signs off.
The Catalogue in Your IDE: Native MCP ServerYour IDE queries the catalogue without leaving the editor — and without bypassing a single permission.
Impact Analysis Before Breaking AnythingBefore dropping a column, you see the 14 dashboards and 3 models that will break.