DataStax is making quite a lot of enhancements to its improvement platform that may enable builders to extra simply implement retrieval augmented era (RAG) of their generative AI functions.
“The Generative AI stack is a giant and sophisticated ball of expertise that many are working to get their arms round. We’re targeted on serving to builders keep true to their roots to allow them to do what they do finest: construct and develop, moderately than worrying about utility infrastructure. We’re delivering a cutting-edge, end-to-end stack to make this quite a bit simpler,” mentioned Ed Anuff, chief product officer at DataStax.
The corporate is releasing Langflow 1.0, the latest model of the visible framework for constructing RAG functions, which DataStax acquired again in April. This replace consists of dozens of integrations with different generative AI instruments, together with LangChain, LangSmith, OpenAI, Hugging Face, and Mistral. It additionally features a cloud-hosted model of Langflow inside DataStax Cloud and entry to LangSmith’s observability capabilities.
DataStax additionally introduced RAGStack 1.0. RAGStack is an out-of-the-box resolution that features numerous instruments wanted to implement RAG.
RAGStack 1.0 consists of help for Langflow, a information graph designed for generative AI, ColBERT powered by Astra DB, Text2SQL and Text2CQL, and Vectorize, which is a software that lets builders select their embedding service.
The corporate can be demoing these new updates at its RAG++ occasion in San Francisco tonight.
You may also like…
RAG is the following thrilling development for LLMs
Protected AI improvement: Integrating explainability and monitoring from the beginning
