Milvus SDK Evolution

4 posts · Dec 2021 – Apr 2025

What this trail covers

Milvus SDK v2 introduces a unified API across Python, Java, Go, and Node.js, native asynchronous support, a performance-boosting Schema Cache, and a simplified MilvusClient interface. The RESTful API now offers near feature parity with gRPC. The Python SDK's MilvusClient streamlines collection creation, indexing, and loading into a single operation. The MCP Server is introduced for AI integration, enabling AI agents to interact with Milvus clusters.

All Milvus trails

Apr 2025Introducing Milvus SDK v2: Native Async Support, Unified APIs, and Superior PerformanceMilvus SDK v2 introduces native asynchronous support (e.g., `AsyncMilvusClient` in Python) for improved concurrency in operations like batch inserts and parallel queries. A Schema Cache is implemented to store collection schemas locally, reducing network requests and CPU overhead for high-frequency operations, leading to lower latency and better scaling. The RESTful API now has near feature parity with gRPC, allowing core operations like collection management and indexing via REST. SDKs across languages are standardized around a `MilvusClient` class for a consistent developer experience. The PyMilvus `MilvusClient` simplifies collection creation, index building, and data loading into a single call. The MCP Server is introduced for AI integration, enabling AI agents to interact with Milvus clusters via the Model Context Protocol.
Jul 2023Milvus 2.2.12: Easier Access, Faster Vector Search Speeds, and Better User ExperienceMilvus 2.2.12 introduces RESTful API support, allowing users to access Milvus without installing a client, simplifying client-server operations. It also enables vector retrieval during Approximate Nearest Neighbor (ANN) searches by allowing the vector field to be an output field in HNSW-, DiskANN-, or IVF-FLAT-indexed collections. The `json_contains` function is added for JSON array operations. Performance enhancements include reduced overhead for extensive topk searches, improved write performance in partition-key-enabled and multi-partition scenarios, and optimized CPU usage for large machines. Various issues related to disk usage, compaction, data deletion, and bulk insertion failures are also addressed.
Mar 2022An Introduction to Milvus Python SDK and APIThis post introduces the Milvus Protocol API, detailing the Protocol Buffers files (milvus.proto, common.proto, schema.proto) that define Milvus's RPC interfaces. It then explains the ORM-style API of PyMilvus 2.0, highlighting its characteristics such as operating directly on objects, isolating service logic, and hiding implementation complexity. It also covers connection control and object-based operations for collections, partitions, and indexes.
Dec 2021Get started with Milvus_CLIIntroduces Milvus_CLI, a command-line tool built on the Milvus SDK for Python. It enables users to connect to Milvus, create collections, list/describe collections, import data, query data, and delete collections using interactive commands. It supports all platforms and offers both online and offline installation methods.
Advertisement