Milvus is an open-source vector database designed to help enterprises and developers to unlock the potential of unstructured data through vector similarity search. Since its donation to LF AI & Data in 2020, Milvus has grown into one of the foundation’s most active and fast-evolving projects, thanks to the incredible efforts of its global open-source community. It is purpose-built to bridge the gap between unstructured data and search applications, enabling seamless semantic search, hybrid filtering, and keyword-based capabilities.
As a founding project of the Open Platform for Enterprise AI (OPEA) initiative under the LF AI & Data Foundation, Milvus contributes to accelerating innovation in AI and data through open-source collaboration and community engagement.
Announcing Milvus 2.5: A Milestone in Unified Search
The release of Milvus 2.5 marks a major advancement in unified search by integrating multiple search paradigms—semantic, keyword, and metadata filtering—into a single solution. This update introduces powerful new features that simplify complex search operations while boosting efficiency.
Built-In Full-Text Search
One of the most notable advancements in this release is the introduction of built-in full-text search powered by Sparse-BM25. This feature allows for more sophisticated search systems by combining text matching and vector similarity search. Users can now input raw text directly, letting Milvus handle the complexities of processing it in real-time using built-in analyzers. Whether dealing with technical documentation or general text data, this feature supports both semantic understanding and keyword precision for superior results.
Enhanced Text and Data Processing
- Text Match: Combines natural language processing (NLP) for precise keyword matching, with vector similarity filtering.
- Bitmap Index: Optimizes searches on low-cardinality fields, delivering faster query performance.
- Nullable & Default Values: Offers greater flexibility in managing scalar fields, simplifying data migration and real-world use cases.
Exciting Beta Features
- Cluster Management WebUI: A dashboard for monitoring system health, performance, and potential issues especially useful for managing distributed Milvus deployments.
- Clustering Compaction: Improves data organization within large-scale collections, resulting in faster queries and more efficient storage.
Additional Upgrades
Milvus 2.5 includes several infrastructure improvements:
- Enhanced HNSW support through Faiss integration.
- IPv6 support for modern networking environments.
- CSV bulk import capabilities.
- Performance and security upgrades for storage and logging systems.
Powered by a Global Open Source Community
Milvus’ growth has been fueled by its active and enthusiastic community of contributors who have propelled the project’s innovations forward. Being part of the LF AI & Data ecosystem ensures Milvus benefits from cross-project collaboration and global visibility while continuing to push the boundaries of enterprise AI solutions.
Ready to Explore Milvus 2.5?
Visit milvus.io today to learn more about these exciting updates and get started with the latest version of Milvus. Or explore the Milvus notebook gallery to get inspired. Join the global community that’s redefining what’s possible with search and unstructured data. With Milvus, the future of unified search is in your hands.
LF AI & Data Resources
- Learn about membership opportunities
- Explore the interactive landscape
- Check out our technical projects
- Join us at upcoming events
- Read the latest announcements on the blog
- Subscribe to the mailing lists
- Follow us on Twitter or LinkedIn
