Cloud database company Couchbase Inc. said today it’s trying to reduce the complexity of agentic application development with the launch of a new suite of AI Services that bundle model hosting and data processing capabilities into a single platform.

The company is looking to streamline the development of artificial intelligence-powered “agents” that can automate complex workloads for businesses and accelerate their productivity. These agents have huge potential in the enterprise, but building them is not easy, thanks in part to the complexity and fragmentation of the underlying information technology stacks needed to support them.

Couchbase said it’s providing everything AI agent developers need to get started in a single platform that’s designed for agentic applications. The new Couchbase AI Services offering is generally available to all customers today, and is meant to solve a range of challenges faced by developer teams. For example, enterprises have major concerns about the security and privacy of publicly hosted large language models, the risk of “hallucinations,” the complexity of dealing with multiple development tools, and the struggle to manage data and prompts.

Couchbase Vice President of Product Rahul Pradhan told CRN in an interview that many of the company’s customers have struggled with their AI agent experiments because of what they say is a “fragmented” experience. Their IT stacks typically consist of structured and unstructured data, public LLMs, vector processing tools, development tools and open application programming interfaces – and all of this must be held together by fragile, customized code.

By providing a unified platform with all of these components, Couchbase said, it can eliminate this fragmentation. For instance, it’s integrating its own agentic design tools with Nvidia Corp.’s AI Enterprise software platform, which provides support for NIM microservices and Nvidia’s Nemotron models and can crunch both structured and unstructured information.

Other features include automatic vector creation, storage and search, intelligent agent memory to support contextual interactions between AI agents, and a unified AI agent catalog that supports full traceability and governance controls. The new services also support SQL++-based analysis from directly within AI applications, helping to streamline development at scale. Moreover, there are new governance and validation tools that enable developers to create guardrails for AI agents and ensure whatever actions they take are allowed under the scope of business rules.

Pradhan said the AI Services effectively transform Couchbase’s database into the “memory and context layer” for AI agents while eliminating complexity and reducing data access latency. He added that customers will be able to use the new services to develop new agentic applications more rapidly and then offer additional services around those apps.

Couchbase Chief Product and Strategy Officer Barry Morris said customers need a robust platform in order to move from agentic AI experimentation to full deployment. “With AI Services we provide developers with an end-to-end platform to build intelligent applications that are secure, trustworthy and performant,” he said. “Customers… [will] benefit from a simplified and repeatable process, resulting from AI capabilities being incorporated coherently across the Couchbase product suite.”

Featured image: SiliconANGLE/Freepik AI