Company

Zilliz

Zilliz develops and operates Zilliz Cloud, a fully managed vector database service built on Milvus, offering real-time serving, on-demand search, and batch analytics over vector and lake data. It also maintains the open-source Milvus vector database. Its offerings serve AI application developers building search, retrieval-augmented generation, and recommendation systems.

Status: Active; Checked 29 Oct 2026
Based in: Redwood City, California, United States; Source ↗
Founded: 2017; Source ↗
Tools listed: 2

Tools from Zilliz

Milvus
Open-source vector database for large-scale AI search and retrieval
Data · Install · Checked 29 Oct 2026
Free open-source softwareCompute, model-provider usage and optional managed hosting cost extra.
Towhee
Open-source framework for building embedding pipelines over unstructured data
Coding · Install · Checked 29 Oct 2026
Free open-source softwareCompute, model-provider usage and optional managed hosting cost extra.

@zilliz_universe on X

We read everything Zilliz posts on X and summarize what’s worth your time: launches, new features and price changes. The noise stays out.

Zilliz@zilliz_universeNew featureNov 2View on X ↗

Milvus 3.0 preview adds aggregation, sorting, long text

“We’re excited to share that 𝗠𝗶𝗹𝘃𝘂𝘀 𝟯.𝟬.𝘅 𝗶𝘀 𝗻𝗼𝘄 𝗶𝗻 𝗣𝘂𝗯𝗹𝗶𝗰 𝗣𝗿𝗲𝘃𝗶𝗲𝘄 𝗳𝗼𝗿 𝗗𝗲𝗱𝗶𝗰𝗮𝘁𝗲𝗱 𝗰𝗹𝘂𝘀𝘁𝗲𝗿𝘀 𝗼𝗻 𝗭𝗶𝗹𝗹𝗶𝘇 𝗖𝗹𝗼𝘂𝗱! 🎉 This release brings more of what developers need for rich search experiences into Milvus: server-side sorting and aggregation, long-text support, and more flexible schema changes. Here’s what’s new: • 𝗔𝗴𝗴𝗿𝗲𝗴𝗮𝘁𝗲 𝘀𝗲𝗮𝗿𝗰𝗵 𝗿𝗲𝘀𝘂𝗹𝘁𝘀: Group retained ANN candidates by scalar fields, calculate per-group metrics, and return representative hits in one request. You can also sort search hits by fields such as price or rating. • 𝗦𝘁𝗼𝗿𝗲 𝗹𝗼𝗻𝗴𝗲𝗿 𝘁𝗲𝘅𝘁: Use TEXT for content beyond VARCHAR’s 65,535-byte limit, including text used for BM25 search. • 𝗖𝗵𝗮𝗻𝗴𝗲 𝗲𝘅𝗶𝘀𝘁𝗶𝗻𝗴 𝘀𝗰𝗵𝗲𝗺𝗮𝘀: Add BM25 or MinHash functions to existing VARCHAR fields, along with their generated fields and indexes. You can also drop fields and functions you no longer need. • 𝗦𝗲𝗮𝗿𝗰𝗵 𝗻𝗲𝘀𝘁𝗲𝗱 𝗱𝗮𝘁𝗮: Run vector search on StructArray elements and apply multiple filter conditions to the same element. • 𝗦𝗲𝘁 𝗽𝗲𝗿-𝗲𝗻𝘁𝗶𝘁𝘆 𝗧𝗧𝗟: Use a TIMESTAMPTZ field to expire individual records, or extend their lifetime with an upsert. • And more! To try it, deploy or use a Dedicated cluster running Milvus v2.6.x and select 𝗧𝗿𝘆 𝗣𝗿𝗲𝘃𝗶𝗲𝘄 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀 on the Cluster Overview page. Learn more: docs.zilliz.com/docs/release-n…”