1 required secret

Databricks Vector Search

Semantic search and similarity queries with Databricks Vector Search indexes

Remote Updated 3 hours ago

Official page View catalog entry Back to catalog

Data AnalyticsVector SearchBusiness Intelligence

About

Run semantic search and similarity queries against Databricks Vector Search indexes through Databricks managed MCP servers, with built-in Unity Catalog security.

Features

Semantic Search: Query and search through your indexed data using Databricks Vector Search
Vector Similarity: Find similar items based on vector embeddings
Index Management: List and access available vector search indexes
Serverless Compute: Leverage Mosaic AI Vector Search pricing model

What you’ll need to connect

Prerequisites

For admins: Ensure the workspace has the Managed MCP Servers preview enabled. See Databricks documentation for your cloud provider.
For users: Ensure your account has access to the Unity Catalog resources (catalogs, schemas) and vector search indexes you want to use.

Setup Steps

Navigate to your Databricks workspace
Go to Catalog → Select your catalog → Select your schema
Verify you have vector search indexes in the selected schema
Create a Personal Access Token in Settings → Developer → Access tokens
Configure the environment variables with your specific values

Configuration

Environment variables

Variable	Description	Required
`DATABRICKS_WORKSPACE_URL`	Databricks workspace hostname, (e.g., https://your-workspace.cloud.databricks.com)	Yes
`DATABRICKS_VECTOR_SEARCH_CATALOG`	Vector search catalog,	Yes
`DATABRICKS_VECTOR_SEARCH_SCHEMA`	Vector search schema, you can find it under the catalog	Yes

Request headers

Header	Description	Required
`Authorization`	Databricks PAT	Yes

Remote runtime

Url Template: ${DATABRICKS_WORKSPACE_URL}/api/2.0/mcp/vector-search/${DATABRICKS_VECTOR_SEARCH_CATALOG}/${DATABRICKS_VECTOR_SEARCH_SCHEMA}