milvus-io/pymilvus

PyMilvus is a Python SDK for Milvus Vector Database that enables developers to interact with Milvus instances from Python applications.

The tool solves the problem of integrating vector database functionality into Python workflows by providing a client library that communicates with Milvus servers. It handles the protocol layer between Python applications and the Milvus backend, allowing developers to perform vector operations, manage collections, and execute similarity searches without directly managing low-level networking or serialization concerns.

Developers should adopt PyMilvus if they are building applications that require vector similarity search and are already using or planning to deploy Milvus. The SDK maintains compatibility across multiple Milvus versions, with specific recommended versions documented for each Milvus release line. It supports Python 3.9 and later, and installation is straightforward via pip. The project includes optional dependencies for model-related functionality, allowing developers to install only what they need.

The project maintains active development with regular updates aligned to Milvus server releases. Development uses the uv package manager for local workflows, and the team enforces code quality through pre-commit hooks that automatically run linting and formatting checks. Test suites are maintained and documented for contributors. The project publishes development builds to Test PyPI from development branches, allowing early testing of unreleased features before official releases. Documentation is kept current and available online, with contribution guidelines established for those wishing to participate in development.