This month, we welcomed the general availability of Pinecone serverless on AWS, the public preview of Private Endpoints for AWS PrivateLink, and a fast-growing list of new integrations. Check out some of our product highlights below or read the full release notes.
Pinecone serverless reinvented vector databases for fast, accurate vector search at any scale with up to 50x lower costs. Serverless is currently available on AWS in us-west-2, us-east-1, and eu-west-1 regions, with more regions and support for Azure and GCP coming later in the year. Pinecone serverless also comes with:
Join developers from more than 20,000 organizations like Gong, Notion, and Shortwave and start building on serverless for free. If you’re already using pod-based indexes, you can migrate your data to serverless indexes for free.
Enterprise users can now easily enable Private Endpoints for AWS PrivateLink. In addition to encryption at rest and in transit, Private Endpoints provides additional security by ensuring that your data traffic traverses the AWS network without public internet exposure.
Enabling Private Endpoints for AWS PrivateLink
Private Endpoints also reduces the risk of exposing your VPC resources to the Internet (or other outside networks) through a misconfiguration, and minimizes the risk of unauthorized access to your Pinecone indexes.
The Java client (v1.2.0) now supports list record IDs in a namespace and with a common ID prefix. ID prefixes enable you to query segments of content, which is especially useful for managing RAG applications where you often need to chunk large documents into smaller segments. This feature is also supported in the Node Js (v2.1.0 and later) Python client (v3.1.0 and later).
Our Python client (v4.0.0 and later) improves your vector upsert throughput by 3x. This is a breaking change if you use the optional GRPC addon (installed with pinecone-client[grpc]).
We’re rapidly expanding our ecosystem to help you streamline your AI application development. You can now access the Pinecone Copilot Extension through our GitHub Marketplace listing to get personalized recommendations. Get started now with this demo.
Other new integrations you can use include data sources Apify, Estuary, Flowise, Unstructured, and StreamNative, frameworks like Contex Data and OctoAI, models like Jina AI and Voyage AI, and observability tools such as Traceloop.
Check out the release notes for a running list of all product and feature releases. If you’re new to Pinecone, try it out for free today.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。