High-performance open-source vector database Qdrant has added new enterprise-level security and management tools to its cloud offering, allowing companies to deploy and scale up artificial intelligence applications.

With today’s announcement, Qdrant Cloud, the company’s managed cloud-based vector database, has added simplified management for clusters, cloud role-based access controls, database application programming interface keys for granular control, advanced monitoring and observability updates alongside cloud API automation.

Vector database technology is core to scaling up AI deployments because it acts as data and knowledge storage for AI models and apps, where data is kept as vectors, which is crucial for modern AI. That allows for the generation and scaling of semantic search, recommendation systems and image recognition where the relationship between concepts can be represented based on their meaning.

These databases enable the production of chatbots and AI agents capable of performing actions based on user intent. They essentially provide “memory” and understanding that allow AI to move beyond simple keyword-based interaction and engage in complex, context-aware actionable conversation.

“AI agents rely on real-time, high-dimensional vector search to retrieve relevant information, enabling complex, multi-step decision-making across dynamic environments,” André Zayarni, Qdrant’s co-founder and chief executive, told SiliconANGLE. “Qdrant’s vector database provides the foundation for agentic applications with multimodal data retrieval, and high-throughput vector processing, enabling adaptive learning workflows. This enables agents to reason over diverse inputs, refine their knowledge, and operate efficiently at scale.”

A new Cloud API now provides simplified management by allowing users to programmatically manage clusters, authentication and deployments using cloud configurations and automate deployment scaling with minimal overhead. The system also gives cloud architecture engineers to use infrastructure-as-code tools such as Terraform to recreate version-controlled deployments more easily. Qdrant said these updates will save time and cost for cloud operations teams setting up and tearing down AI deployments.

Fine-grained authentication and permission controls for cluster management are now available through Cloud role-based access controls, billing and hybrid deployment in early access. Database API keys can be used to provide per-cluster and per-collection access, restricted by read-only or read/write permissions per key. The company said this fine-grained access will provide enterprise-grade functionality for application security and control between different roles allowing greater flexibility for AI agents to do their work.

Single sign-on is also available through several providers, including Okta, Google Workspace, Azure AD (Entra ID) and PingFederate – or through Security Assertion Markup Language to connect another third-party authentication provider.

New real-time monitoring has been added with integration with OpenMetrics-based data monitoring formats that allow engineers and operations teams to easily track telemetry for query performance, latency, request volumes, processor usage, memory usage and disk space. The company said it’s compatible with DataDog, Grafana and numerous other monitoring tools.

Image: SiliconAngle/Microsoft Designer