OceanBase Open-Sources AI-Native Database seekdb

OceanBase has launched and open-sourced seekdb, an AI-native database designed for emerging AI applications, announced in a press release. The company introduced seekdb at its 2025 Annual Conference, highlighting its unified approach to handling vector, text, structured, and semi-structured data in a single engine.

Seekdb enables hybrid search capabilities that combine vector search, full-text retrieval, and scalar filtering within one query. It supports real-time data writes with full ACID compliance and maintains compatibility with the MySQL ecosystem. The system can deliver millisecond-level responses on billion-scale multimodal datasets.

The database is lightweight, requiring only one CPU core and 2 GB of memory. It supports one-command installation via pip and offers embedded and client/server deployment options. Seekdb integrates with over 30 AI frameworks and Model Context Protocols, including HuggingFace, Dify, and LangChain, and provides both SQL and Python SDK interfaces.

Seekdb is open-sourced under the Apache 2.0 license and is available on GitHub, allowing developers to modify and extend its functionality. OceanBase stated that the database is built to simplify the development of AI applications and improve access to multimodal data for enterprise and research use.