Selectel has expanded its cloud database offerings by launching a managed PGVector service. This solution is designed to handle vector data, enabling semantic search, recommendation engines, and RAG-based AI systems. The service is built on PostgreSQL with the pgvector extension pre-installed, eliminating the need for manual configuration.
The company manages the installation, maintenance, and operational support of the database. According to Alexander Grishin, head of data storage product development at Selectel, the service is intended to accelerate project deployment and optimize maintenance costs. Users can scale the infrastructure based on current load and pay only for actual usage.
The PGVector service offers high availability starting from a two-node configuration. Selectel reports that this setup reduces costs for fault tolerance by up to 33% compared to standard PostgreSQL configurations. Furthermore, the solution supports multi-zone (Multi-AZ) clusters to ensure continuous operation during site failures.
The new service is now available through the Selectel control panel. It integrates with the company's existing product ecosystem, allowing teams to connect the database to LLM frameworks for building corporate knowledge base assistants and other AI-driven applications.
