Skip to content

Feature Request: Add Milvus Vector Store Integration for RAG #4420

@zc277584121

Description

@zc277584121

Description

Add Milvus as a vector database option for knowledge base RAG workflows, following the existing Spanner integration pattern.

Motivation

ADK currently supports Vertex AI RAG, LlamaIndex, and Spanner for vector-based retrieval. Milvus is a widely adopted open-source vector database purpose-built for AI applications. Adding Milvus support would give developers a lightweight, self-hosted (or cloud-managed via Zilliz) alternative for building RAG agents.

Proposed Solution

  • MilvusVectorStore: Utility class for collection setup, data ingestion, and similarity search.
  • MilvusToolset: BaseToolset implementation exposing a similarity_search tool to LLM agents.
  • MilvusToolSettings / MilvusVectorStoreSettings: Pydantic configuration classes.
  • Users provide their own embedding function (e.g., Google GenAI gemini-embedding-001).
  • pymilvus>=2.5.0 as an optional dependency (google-adk[milvus]).

Related PR

#4417

Metadata

Metadata

Assignees

No one assigned

    Labels

    services[Component] This issue is related to runtime services, e.g. sessions, memory, artifacts, etc

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions