Catalog

Vector fields

Name: Nexoxa Bridge
Author: Nexoxa

pgvector columns and similarity helpers.

The story

These pages use illustrative scenarios to walk through real workbench panels. Sidebar labels, Bridge paths, and connection schemes match the app; characters and timelines are examples for learning.

Embedding search slow, opens Vectors — pgvector columns, dimensions, index types — tunes HNSW params with evidence.

The Catalog panel Vectors is where Bridge turns raw metadata into decisions — counts, definitions, dependencies, and links into the grid or Query lab. No information_schema archaeology; no guessing which mat view is stale.

Open Vectors from the left sidebar after you connect; everything here reflects the live connection, not a cached export from yesterday.

Bridge path: /database/vectors

Step 1

Vector columns

embeddings vector(1536) on articles — dimension matches model. Wrong dim caught early.
Index type listed.

pgvector columns and similarity helpers. /database/vectors. pgvector metadata.

Step 2

Similarity helpers

Test similarity query in lab — cosine vs L2 choice informed by column index.
Latency acceptable.

Query patterns in Query lab. Extension requirement in Extensions.

Step 3

Extension dependency

vector extension version in Extensions panel — upgrade planned with Postgres minor.
No surprise DROP.

Link to Catalog → Extensions.

This panel appears in the Nexoxa Bridge left sidebar after you connect with the matching connection URL. See the connection docs above for postgresql://, ssh://, sftp://, s3://, or smtp:// examples.

Related connections

Database connection URLs