Hi friends,<p>Excited to share our latest project, pgvector-remote, with you all!<p>pgvector-remote is an extension designed to effortlessly incorporate remote vector stores, like Pinecone, into pgvector. With pgvector-remote, harnessing the power of a remote (Pinecone) index becomes as easy as utilizing ivfflat or hnsw.<p>Why pgvector-remote?<p>1. Simplifies integration of memory-intensive vector indexes like Pinecone into standard database servers.<p>2. Addresses critical feature gaps in pgvector, such as metadata filtering for AI applications.<p>3. Hides the complexity of maintaining and synchronizing vector stores from the user.<p>Key Features:<p>+ Achieve <10ms latency and 200QPS throughput, similar to Pinecone.<p>+ Enjoy metadata filtering capabilities, just like Pinecone.<p>+ Benefit from SQL compatibility and transaction support akin to PostgreSQL.<p>Usage:
CREATE INDEX ON items USING PINECONE (embedding);<p>We're continuously enhancing pgvector-remote and would love to hear your feedback!