Couchbase on Monday introduced Capella AI Services, a set of new services under development aimed at making it easier and faster for customers to develop generative AI applications, including agents that autonomously take on some of the work previously performed by humans.
The suite, now in private preview, includes features such as model services to connect data with large language models (LLMs) and unstructured data services to transform unstructured text documents into JSON data.
General availability is planned for some time in 2025, according to Matt McDonough, Couchbase's senior vice president of product and partners.
Enterprises are continuing to increase their investments in developing AI tools such as conversational assistants and agents that let users interact with data using natural language rather than code. Given the rising interest in developing AI tools, Couchbase's development of Capella AI Services provides features that are of interest and importance to customers, according to Stephen Catanzano, an analyst at TechTarget's Enterprise Strategy Group.
"It is significant because it unifies AI model hosting, data processing and development tools into one platform, reducing complexity, latency and costs," he said. "It enhances data security and governance while accelerating time-to-market for AI applications, addressing major challenges in AI development."
Based in Santa Clara, Calif., is a NoSQL database vendor first known as Membase before a 2011 merger with CouchOne. Capella, a database-as-a-service platform for cloud-based customers, was first launched in 2021. Couchbase Enterprise, meanwhile, is the vendor's platform for on-premises users.
Couchbase competitors include other database specialists such as Redis and MongoDB, as well as tech giants such as AWS, Google, Microsoft and Oracle that offer database platforms.
A new suite for AI
Enterprise interest in developing AI applications has surged in the two years since OpenAI's November 2022 launch of ChatGPT marked a significant improvement in generative AI technology.
Now, using generative AI models in conjunction with their proprietary data, organizations can build applications that enable users to interact with data using natural language rather than code. This for the first time, enables the widespread use of analytics to inform decisions. In addition, by combining generative AI capabilities with proprietary data, organizations can automate certain repetitive, time-consuming processes to make data experts more efficient.
Many data management and analytics vendors have created AI development environments within their platforms to make it easier for customers to build applications that make employees smarter and more efficient.
For example, Couchbase competitors MongoDB and SingleStore now provide AI development capabilities. Similarly, data platform vendors Databricks and Snowflake, along with tech giants AWS, Google Cloud and Microsoft, all offer AI development environments.
Now, Couchbase is following suit with the introduction of Capella AI Services.
With AI becoming a critical tool for enterprises, customer and partner feedback played a role in Couchbase's decision to develop Capella AI Services, according to McDonough. So, too, did the vendor's observation of market trends, including rising interest in agentic AI in recent months.
"Developers currently lack well-defined best practices for building and deploying agentic applications," McDonough said. "With the new services, we're addressing that."
In particular, Couchbase wants to enable its customers to be able to make use of unstructured data when building AI tools, he continued.
Unstructured data such as text, images and audio files now account for more than 80% of all data. Its unstructured nature, however, makes it difficult to discover and operationalize in a data lake or database housing massive amounts of data. Complex processes such as vectorization and retrieval-augmented generation (RAG) can structure unstructured data to make it actionable. As a result, they have become common elements of AI development environments and are included in Capella AI Services.
"We want to securely and effectively equip developers with AI so they can take advantage of its greatest strength, which is to process unstructured data and turn it into something meaningful," McDonough said.
Capella AI Services includes the following features:
Model Services, which provides access to LLMs and embedding models as well as capabilities that support RAG pipelines such as prompt and conversion caching and keyword filtering.
Unstructured Data Services to clean unstructured documents and transform them into JSON data to prepare them for the vectorization that makes them discoverable.
Vectorization Services to automate vectorization and indexing of unstructured data stored in Couchbase's Capella database.
AI Agent Catalog Services, a centralized storage location for tools, governance capabilities, lineage, metadata, model input prompts and audit information aimed at speeding agentic AI application development.
Capella AI Functions to enable developers to embed AI-powered analysis into application workflows without having to use external tools or write customized code.
Some of the features included in Capella AI Services, such as support for vector search and storage and transformation of unstructured data to JSON, were previously available as standalone tools. The suite repackages them as services and combines them with additional features such as model services and AI agent catalog services.
Together, the capabilities that make up Couchbase's Capella AI Services are on par with what competing vendors are offering, according to Catanzano.
"Capella AI Services is competitive, with features tailored to enterprise-grade AI needs like RAG pipelines and data security," he said. "It may still trail [tech giants] like AWS or Azure in breadth of AI model options and ecosystem integrations but excels in ease of use and developer-centric tools."
Highlight features, meanwhile, include the data catalog capabilities in AI Agent Catalog Services and Capella AI Functions, Catanzano continued.
"These features stand out for simplifying development and enhancing traceability and productivity," he said.
Next steps
With the features in Capella AI Services now in private preview, Couchbase's product development plans for 2025 include a continued focus on enabling developers to build AI applications, according to McDonough.
It is significant because it unifies AI model hosting, data processing and development tools into one platform, reducing complexity, latency and costs. It enhances data security and governance while accelerating time-to-market for AI applications, addressing major challenges in AI development.
Stephen CatanzanoAnalyst, Enterprise Strategy Group
"We're focused on helping developers take advantage of AI in many different ways, whether they're just getting started with basic search features or building advanced AI agent applications that can independently take the best course of action based on human-set goals," he said.
Included in that are plans to upgrade vector search capabilities, expand multimodal AI capabilities from the cloud to the edge to capture data from more sources and improve Capella AI Services so developers can build applications without having to use multiple platforms.
In addition, he suggested that Couchbase could add data observability capabilities to help customers improve the performance of AI tools and should focus on broadening its ecosystem to attract potential new customers. Increasing the number of LLMs and embedding models with which it integrates enables developers to better select the model that best fits their needs while adding integrations with more AI development platforms to provide users access to their preferred environments.
"Couchbase should focus on expanding model integrations and providing advanced analytics and monitoring tools to help organizations manage and optimize AI applications at scale," Catanzano said. "Enhanced interoperability with third-party AI ecosystems would also strengthen its appeal."
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.
