OceanBase Unveils and Open-Sources AI-Native Database seekdb at 2025 Annual Conference

• Unifies hybrid search for vector, text, structured and semi-structured data in a single engine
• Enables in-database AI workflows and millisecond response on billion-scale multimodal data

BEIJING, Nov. 18, 2025 /PRNewswire/ -- At its 2025 Annual Conference today, OceanBase officially launched and open-sourced seekdb, its first AI-native database designed specifically for emerging AI applications.

OceanBase seekdb brings high-performance hybrid search by unifying vector, text, structured and semi-structured data in a single engine. With OceanBase seekdb, developers can rapidly build AI applications with as few as three lines of code and capable of delivering millisecond-level search responses in billion-scale multimodal datasets.

According to Gartner, spending on databases integrating generative AI capabilities is projected to surge from 65 billion USD in 2025 to 218 billion USD by 2028, accounting for 74% of the total market in that year (Gartner Forecast Analysis: GenAI in Database Management Systems). Meanwhile, a recent report from the MIT Media Lab indicates that 95% of enterprise generative AI pilots deliver no measurable return, due to fragmented multimodal data, overly complex system, and difficulties in access control (The GenAI Divide, State of AI In Business 2025).

OceanBase seekdb is purpose-built to address these challenges, with the following key features:

AI-native Capabilities for Hybrid Search

OceanBase seekdb supports the fusion of vector search, full-text retrieval, and scalar filtering within a single query, employing a multi-stage retrieval mechanism to achieve higher accuracy with low latency. Built on a proven transactional engine, it enables real-time data writes with full ACID compliance and seamless compatibility with the MySQL ecosystem. Additionally, OceanBase seekdb provides unified storage and retrieval for multimodal data types, including scalars, vectors, text, JSON, and GIS.

Lightweight and Easy-to-Use

OceanBase seekdb requires as little as 1 CPU core and 2 GB of memory to run. It supports one-command installation via pip install, starts up in seconds, and offers two deployment modes, including embedded and client/server, making it easy to integrate into AI agents, toolchains, and local applications, significantly lowering the barrier to AI development.

Open-Sourced and Developer-Friendly

OceanBase seekdb is open-sourced under the Apache 2.0 license from day one and is publicly available on GitHub, allowing developers to freely use, modify, and extend its codebase. Moreover, it seamlessly integrates with over 30 mainstream AI frameworks and Model Context Protocols, including HuggingFace, Dify, and LangChain, and provides SQL and PythonSDK to support diverse developer preferences.

Evan Yang, Chief Executive Officer of OceanBase, noted: "As AI transitions from cutting-edge innovation to real-world applications, the bottleneck lies not only in models but also in data handling. Future-oriented databases must serve the needs of both human beings and intelligent agents, supporting transactional, analytical, and AI workloads within a single system to drive real-time, trustworthy intelligence directly from the data side. OceanBase is dedicated to exploring a paradigm shift in how databases support the AI era."

Over the past several years, OceanBase has supported AI application deployment across multiple industries. Leading enterprises, including Lalamove, China Unicom, and Trip.com, have adopted OceanBase to build stable, high-performance RAG and search-augmented systems, empowering higher-quality search and question-answering experiences.

Since its commercial launch in 2020, OceanBase has been adopted by more than 4,000 customers worldwide across industries including financial services, internet, public services, telecommunications, and retail, achieving an average annual growth rate of over 100% for five consecutive years. Today, OceanBase delivers services across more than 240 availability zones in over 60 geographic regions. 

About OceanBase

OceanBase is a distributed database launched in 2010. It provides strong data consistency, high availability, high performance, cost efficiency, elastic scalability, and compatibility with mainstream relational databases. It handles transactional, analytical, and AI workloads through a unified data engine, enabling mission-critical applications and real-time analytics.

To learn more, please visit: https://www.oceanbase.com/

source: OceanBase