←
Pinecone2,613 posts
Pinecone avatar

Pinecone

@pinecone

Pinecone is the trusted AI knowledge company. Pinecone's mission is to make AI knowledgeable.

◷ Joined July 2020▧ 601 Photos and videos
Pinecone avatar
Pinecone@pinecone·Aug 6

Nexus is GA. Enterprise knowledge, compiled once, query-ready for agents. On Sierra's τ-Knowledge: GPT-5.2 +12% accuracy, -80% cost. GPT-5.5 same accuracy, -77% cost. Read the full announcement: pinecone.io/blog/pinecone-ne…

1:15
Joerg avatar
⟳ Pinecone retweeted
Joerg@joerg_schad·3h

Questions or cool ideas around vector search, full-text search, pg_pinecone (postgres integration), or AI in general? Join us at the Pinecone office hours starting in 30 minutes! 🚀 discord.com/events/138365843… @pinecone #postgres

Pinecone avatar
Pinecone@pinecone·4h

The pg_pinecone Postgres extension now deploys on @Railway in one click. In benchmarks on 1 million vectors, pg_pinecone's index was about 15 times smaller than a pgvector HNSW index using halfvec, so it costs less to run on Railway. 🔗 pinecone.io/blog/pg-pinecone…

Pinecone avatar
Pinecone@pinecone·7h

Your answer is somewhere in a huge pile of documents. But which ones? Ashwin Padaki explains how embeddings turn that into a nearest neighbor search.

1:10
Pinecone avatar
Pinecone@pinecone·Nov 17

Vector-to-document index conversion is now available. Teams get full-text search on the data already in Pinecone without reimporting. One conversion job builds a new document index while the legacy vector index keeps serving traffic. 🔗 pinecone.io/blog/index-conve…

Pinecone avatar
Pinecone@pinecone·Nov 12

Through the end of the year, the Builder plan gets twice its included tokens for Gemini Embedding 2: 20M a month.

Pinecone avatar
Pinecone@pinecone·Nov 12

You can now use Google's Gemini Embedding 2 with Pinecone integrated embedding. Set gemini-embedding-2 on a text field and send plain text. That field can serve full-text search too, so hybrid search doesn't need a second index or embedding pipeline. 🔗 pinecone.io/blog/integrated-…

Pinecone avatar
Pinecone@pinecone·Nov 11

We're at AI Engineer Code Summit! Come find the Pinecone team at Booth #4 in the Golden Gate Ballroom. Bring your questions, tell us what you're building, and let's talk about your next steps.

0:08
Pinecone avatar
Pinecone@pinecone·Nov 10

Through the end of the year, Starter and Builder plans get twice their included tokens for OpenAI's models: 10M a month on each model for Starter and 20M a month on each model for Builder.

Pinecone avatar
Pinecone@pinecone·Nov 10

You can now use OpenAI's text-embedding-3 models with Pinecone integrated embedding. Set either model on a text field and send plain text. That field can serve full-text search, so hybrid search doesn't need a second index or an embedding pipeline. 🔗 pinecone.io/blog/integrated-…

Pinecone avatar
Pinecone@pinecone·Nov 10

Pinecone’s support agent doesn’t just search the docs. It pairs RAG with live customer data (monitoring, usage, billing) to answer based on your actual account. Watch Zeke Emerson explain on how Pinecone built it: youtu.be/nvGaAgGH488

Post attachment
Pinecone avatar
Pinecone@pinecone·Nov 9

See you tomorrow, San Francisco! The Pinecone team will be at Booth #4 for AI Engineer Code Summit in SF this week. Stop by to meet the team, get a demo, and talk through what you're building with agents. The Opening Reception is Tuesday from 5:30 to 8:30 PM PT. We will be at the expo Wednesday and Thursday (Nov 11-12) as well.

0:08
Pinecone avatar
Pinecone@pinecone·Nov 6

How do you find nearest neighbors without a graph? Jörg Schad, VP of Engineering at Pinecone, explains how it works. pg_pinecone's native engine uses IVF + RaBitQ: split vectors into clusters, search only the closest one, and filter as you go. Try it: github.com/pinecone-io/pg_pi…

1:22