Yugabyte, the distributed SQL database expert, is unveiling extended vector search functionality and PostgreSQL compatibility for YugabyteDB, the AI-ready, multi-modal, distributed PostgreSQL database for cloud native apps. This release is designed to provide developers with the familiarity of PostgreSQL while innovating the flexibility and tools they have at their disposal for creating powerful AI applications at scale.
YugabyteDB, a 100% open source database, retains the power of PostgreSQL while underpinning it with a cloud native architecture—enabling it to run everywhere, rapidly scale across distributed clusters, and effortlessly withstand any hardware failures.
Key to this release is Yugabyte’s attention to both customer needs and industry trends, which both point to the need to run effective, future-proof generative AI (GenAI) and AI-powered applications. With more developers adopting Postgres as the standard method for creating AI apps, Yugabyte works to alleviate the various architectural complexities associated with Postgres, such as scale, availability, operational simplicity, and more.
As part of this commitment, Yugabyte is announcing new vector search, PostgreSQL, and multi-modal functionality to meet the needs of developers within a single, ultra-resilient distributed database, according to the company. New capabilities debuted in this update include:
- YugabyteDB MCP Server for seamless AI-powered experiences in applications
- Newfound support for LangChain, OLLama, LlamaIndex, AWS Bedrock, and Google Vertex AI
- Multi-modal API support with the addition of MongoDB API support for scaling MongoDB workloads in addition to PostgreSQL (YSQL) and Cassandra (YCQL)
- Online upgrades and downgrades across major PostgreSQL versions with zero downtime
- Enhanced PostgreSQL compatibility with generated columns, foreign keys on partitioned tables, and multi-range aggregates
- Built-in Connection Pooling that can support tens of thousands of connections per node
At its core, these capabilities enable Yugabyte to “take the simplicity of Postgres [and] give people the API or interface they're used to, with all of the complexity taken care of at the architectural level,” said Karthik Ranganathan, co-founder and CEO, Yugabyte. Then, “We take a unifying and spanning approach, where we say, ‘All of these different independent pieces that are integrated with Postgres, we're going to bring them together and make them work together.’”
The YugabyteDB MCP Server—alongside support for LangChain, OLLama, LlamaIndex, AWS Bedrock, and Google Vertex AI—works to establish a cohesive plane from which the user’s AI ecosystem and the data platform can work together. Multi-modal API support across YSQL, YCQL, and MongoDB, workloads, as well as vector indexing and search with the YugabyteDB MCP server, empowers enterprises to build and deploy highly resilient, "ready-to-scale" retrieval-augmented generation (RAG) and AI-powered applications with 99.99% or higher uptime—without the complexity of managing AI workloads.
Additionally, with seamless online upgrades and downgrades, as well as enhanced PostgreSQL compatibility, users benefit from new Postgres features without impacting existing applications—with the ability to revert to an older version if needed. This capability speaks to the way enterprises “all [want to] double down on Postgres as the standard, but they don't want major variants to Postgres. They don't want something that looks so different that you move to that database, which claims it's post Postgres, but it's a one way street—and you cannot come back,” said Ranganathan.
“We've built a framework that unlocks all of [the enhanced Postgres] value, resilience, scale, operational simplicity, cloud native, but knows how to go from one version of Postgres to a major future version of Postgres. And if things don't work, you can come back to the old version,” Ranganathan continued. “We are truly unique in this aspect compared to anybody else in the industry, because it's ultimately blending true Postgres compatibility with true cloud native, and it's not taking half measures on either side.”
Fundamentally, this release reflects how, “We [Yugabyte] listen to what customers are doing and what they're trying to accomplish, and try to aid in any way possible, as opposed to just selling them something,” noted Ranganathan. “We just want to do what is useful and truly cutting edge.”
To learn more about YugabyteDB, please visit https://www.yugabyte.com/.