DataStax, the company that powers generative AI applications with real-time, scalable data, is debuting the Data API (formerly known as the JSON API)—a document-centric API that forwards a developer-centric experience optimized for vector development and retrieval augmented generation (RAG) use cases—for AstraDB. By providing all the data and complete RAG stack necessary for creating generative AI (GenAI) apps, DataStax’s Data API is suited for any modern application use case that utilizes document-based data, according to the company.
A large part of DataStax’s focus is improving the developer experience from end-to-end, both in regard to the user experience and the cloud, explained Ed Anuff, chief product officer at DataStax.
With this in mind, the Data API brings the power of Apache Cassandra to JavaScript and Python developers, underpinned by an intuitive experience for AI development without the need for extensive Cassandra knowledge.
The Data API is a clear response to the explosive field of AI and GenAI app development, where, according to Anuff, “about half of the signups we get on our cloud service are people that are doing something around vector databases and GenAI.”
By incorporating Cassandra’s zero-delay indexing—which means that as soon as the data is ingested, it’s already indexed, making vector data immediately available for retrieval—the Data API empowers a sense of ease around applying that technology to create powerful GenAI apps.
The Data API also features out-of-the-box support for major GenAI partners—including LangChain, OpenAI, Vercel, Google’s Vertex AI, AWS, and Azure.
“It's very important to us, as we thought about building these APIs, that we did so in a way that optimized the experience and made it very simple for developers,” said Anuff. “We want to make sure that we allow people to move fast, to build [GenAI] quickly…because we're all learning how to build them now together. But we want to also be the company that makes it possible to put those into production at scale, in a cost-effective way.”
Whether developers have previous experience with Cassandra or not, the Data API can facilitate rapid AI and GenAI app development for developers dealing with document-centric data—all with up to 20% higher relevancy, 9x higher throughput, and up to 74x faster response times than other vector databases, according to the company.
“One of our design goals with the Data API was to not have developers have to learn a whole lot of database theory when they're really trying to retrieve content via vector,” explained Anuff. “DataStax has made Cassandra really easy to use [with AstraDB] ...the Data API [also] makes all of this power very easy to use…and [you'll] never have to worry about scale and availability challenges because… [you’ve got] the power of Cassandra behind all that.”
In addition to the launch of the Data API, DataStax is unveiling a series of updates to the developer experience for AstraDB, the vector database for building production-level AI applications. These updates improve the developer portal—which now delivers a Data Explorer capability, new relevancy visualizations, and different portal views, including tabular or document-oriented—which further complements the Data API.
"The seamless user experience, native language client support, and effortless integration with GenAI ecosystems empower us to navigate the complexities of vector search with unprecedented simplicity," said Mark Hauptman, chief executive officer at ReelStar. "With Astra DB, we're not just shortening the development journey; we're accelerating innovation and efficiency in our mission to redefine the landscape of video sharing with GenAI.”
To learn more about AstraDB and its new Data API, please visit https://www.datastax.com/.