Agent Retrieval (Preview) remote MCP server
Query Agent Retrieval (Vector Search 2.0) collections and indexes and run searches
What it does
A Google-managed remote MCP server. Query Agent Retrieval (Vector Search 2.0) collections and indexes and run searches. Aimed at RAG and agent builders, it lets an agent list collections, inspect indexes, search data objects, run batch searches through an HTTP endpoint on Google infrastructure, governed by IAM and audit logging. Currently in preview.
What an agent can do with it
- list collections
- inspect indexes
- search data objects
- run batch searches
- aggregate results
How to set it up
Enable the MCP server for your project and point a remote-capable MCP client at https://vectorsearch.googleapis.com/mcp, authenticating with OAuth or Google credentials.
Full steps: Google's documentation.
Use cases that use it
Low-stock reorder alerts from Shopify inventory
A daily check of Shopify stock levels against sales pace that flags SKUs running low and drafts a… Read →
E-commerce ops · Shopify · Google Sheets · ClaudeReturn and RMA request triage
Sort incoming return requests by reason, eligibility, and urgency, then draft a policy-aligned… Read →
E-commerce ops · Shopify · Gorgias or Zendesk · GmailProduct description drafts from supplier spec sheets
Turn supplier spec sheets into first-draft product titles, descriptions, and bullet points in your… Read →
E-commerce ops · Claude · Google Drive · AirtableProactive shipping-delay customer notices (drafts only)
Spot orders that are late to ship or stalled in transit and draft a short, honest heads-up message… Read →
E-commerce ops · Shopify · Carrier tracking export or webhook · ClaudeSources
Descriptions are written by Agents Lookup unless marked as quoted. Each link opens the original.
- MCP documentation: docs.cloud.google.com/gemini-enterprise-agent-platform/build/vector-search-2/use-agent-retrieval-mcp (checked 2026-11-19)
- Google list where it appears: docs.cloud.google.com/mcp/supported-products