Can I use Vertex AI Search with my own embeddings?

No. Vertex AI Search manages its own embedding pipeline — you upload documents and it handles everything. If you need to supply your own embeddings, use Vertex AI Vector Search instead.

When should I set location to a regional value?

Most unstructured data stores use "global" (the default). Some enterprise features (EU data residency, private endpoints) require a regional location such as "us" or "eu". Check your data store configuration in the Cloud Console.

Is Vertex AI Search the same as Agent Builder?

No. Vertex AI Search (Discovery Engine) is the retrieval layer. Agent Builder and Vertex AI Conversation are higher-level products that bundle search with conversational UI. Gateco connects to the Discovery Engine REST API at the search layer.

Vector DBTier 2

Vertex AI Search

Policy-gated enterprise search on Vertex AI Search — keyword, semantic, and hybrid retrieval with Gateco access control.

Search capabilities

Vector (ANN)	Keyword (ranked)	Keyword (filter)	Hybrid (native)	Grep (substring)	Grep (regex)

<200ms p95 typical (managed cloud, includes Google's internal processing). Policy overhead <25ms p95.

sidecar metadatainline metadata

Vertex AI Search (Discovery Engine) is Google Cloud's managed enterprise search service for unstructured documents, structured data, and websites. It handles ingestion, embedding, indexing, and hybrid search automatically. Gateco connects to it as a Tier 2 retrieval connector.

Unlike Vertex AI Vector Search, Vertex AI Search does not accept raw embedding vectors from clients. It manages its own embedding pipeline. The vector: false flag in search capabilities reflects this — you cannot send query embeddings directly to this connector.

Hybrid search uses Vertex AI Search's internal RRF fusion. Gateco's alpha parameter does not map to a Vertex-side control — when alpha ≠ 0.5, a warning is emitted but native RRF proceeds unchanged.

Vertex AI Search uses the Discovery Engine API (discoveryengine.googleapis.com), not the Vertex AI Platform API. The same GCP service account JSON and auth helper are used for both Vertex connectors.

Sample search config

{
  "project_id": "my-project",
  "location": "global",
  "data_store_id": "my-data-store",
  "serving_config_id": "default_serving_config"
}

Sample policy

{
  "name": "Enterprise search access",
  "effect": "allow",
  "rules": [{
    "conditions": [
      {"field": "principal.department", "operator": "in", "value": ["finance", "legal", "hr"]},
      {"field": "resource.classification", "operator": "in", "value": ["internal", "public"]}
    ]
  }],
  "selectors": [{"connector_type": "vertex_ai_search"}]
}

Policy conditions reference resource.* and principal.* fields. Policy reference →

Frequently asked questions

Can I use Vertex AI Search with my own embeddings?: No. Vertex AI Search manages its own embedding pipeline — you upload documents and it handles everything. If you need to supply your own embeddings, use Vertex AI Vector Search instead.
When should I set location to a regional value?: Most unstructured data stores use "global" (the default). Some enterprise features (EU data residency, private endpoints) require a regional location such as "us" or "eu". Check your data store configuration in the Cloud Console.
Is Vertex AI Search the same as Agent Builder?: No. Vertex AI Search (Discovery Engine) is the retrieval layer. Agent Builder and Vertex AI Conversation are higher-level products that bundle search with conversational UI. Gateco connects to the Discovery Engine REST API at the search layer.

Ready to connect Vertex AI Search?

Follow the step-by-step setup guide or talk to the team for help with your specific configuration.

Setup guide →Talk to the team All integrations