Posted on nosqlgooglealerts. Visit nosqlgooglealerts

The addition of AI services to the company’s Capella cloud database and development platform will provide developers with more control over data, development workflows and AI models.

Couchbase is adding new artificial intelligence capabilities to its Capella cloud database that the company says will streamline the development of agentic AI applications.

The new Capella AI Services, unveiled Monday, provide developers with simplified data integration workflows and more control over data throughout the development lifecycle including putting agentic applications into production, according to Couchbase.

The AI services also help developers mitigate data security and privacy risks by keeping data and large language models – including LLMs running outside of an organization – close together.

[Related: Meeting The Exploding Demand For Data: The 2024 CRN Big Data 100]

With the new Capella AI Services, Couchbase, based in Santa Clara, Calif., is looking to provide data management capabilities needed for the growing wave of AI and generative AI applications development.

“This release is all about our offering, really targeted at developers building AI and agentic or multi agentic applications,” said Matt McDonough, Couchbase senior vice president of product and partners, in an interview with CRN. “There’s a lot of enthusiasm around AI agents, but the industry as a whole lacks well-defined best practices for building and deploying these agentic applications.”

The new AI services come on the heels of Couchbase’s September announcement of expanded columnar and vector search functionality in the Capella database-as-a-service for developing next-generation adaptive applications – including those with AI functionalities. Capella is based on Couchbase’s NoSQL database server.

“The key is [to] make it simple for developers to build, test and deploy AI agents without having to use disparate platforms,” McDonough said. “And do this in a way that reduces latency, operational costs [and] keeping models and data close together as part of this whole agentic AI software development life cycle.”

The new AI services are also an enabler for Retrieval-Augmented Generation workflows that move proprietary data into LLMs, according to the executive.

The new AI services include model hosting, automated vectorization, unstructured data preprocessing and AI agent catalog services – all of which allow developers to prototype, build, test and deploy AI agents. In addition to keeping models and data close together, the services help organizations reduce development complexity, avoid excess latencies and high operational costs often experienced when introducing new technology components and workflows, according to Couchbase.

“The greatest strength of AI is its ability to process unstructured data,” McDonough said. AI agents can take unstructured information, such as a transcription of a meeting, and autonomously incorporate it into operational applications and workflows. But AI agents need flexible databases with the ability to work with complex data types and unstructured data – such as PDF documents and audio files – to be effective, he said.

Couchbase ISV and systems integrator partners will particularly benefit from the new AI Services, McDonough said. ISV partners who develop their applications on the Capella platform can better meet customer requests to add AI agent capabilities to those applications. And global and regional systems integrators can use the new functionality to expand the range of development services they can provide their customers, he said.

The new Capella AI Services include:

• Model services that provide managed endpoints for leading LLMs and embedding models, and capabilities such as prompt and conversation caching, guardrails and keyword filtering to support RAG and agentic workflows.
• Unstructured data services that extract, clean, chunk and transform unstructured documents into JSON, preparing them for vectorization. It also extracts structured information from complex documents and makes it queryable in Capella.
• Vectorization services that automate vectorization and indexing of data stored in Capella.
• AI agent catalog services that accelerate agentic application development by providing a central repository for tools, metadata, prompts and audit information for LLM flow, traceability and governance.
• Capella AI functions that enable AI-driven data analysis directly into application workflows using familiar SQL++ syntax.

The AI services are currently in private preview and are slated to be generally available as part of the Capella cloud database in 2025.