External reviews
45 reviews
from
and
External reviews are not included in the AWS star rating for the product.
fast and easy to setup vector database
What do you like best about the product?
The things I mostly like are:
- that is easy to set up by following the docs
- fast for loading and updating embeddings in the index
- easy to scale if needed
- that is easy to set up by following the docs
- fast for loading and updating embeddings in the index
- easy to scale if needed
What do you dislike about the product?
- that is not open source
- I cannot query the full list of ids from an index (I needed to build a database and a script to track what products I have inside the index)
- customer support by mail takes too much time
- I cannot query the full list of ids from an index (I needed to build a database and a script to track what products I have inside the index)
- customer support by mail takes too much time
What problems is the product solving and how is that benefiting you?
I built a deep learning model for product matching in the ecommerce industry. One of the steps for the system is to find candidates that are potential matches for the searched product. Becase of this, I needed a vector database to store the embeddings (texts and image) for the products for doing a similarity search as a first step of the product matching system.
GWI on Pinecone
What do you like best about the product?
Easy of use and metadata filtering. Pinecone is one of the few products out there that is performant with a query that contains metadata filtering.
What do you dislike about the product?
The pricing doesn't scale well for companies with millions of vectors, especially for p indexes. We experimented with pgvector to move our vectors in a postgres but the metadata filtering performance was not acceptable with the current indexes it supports.
What problems is the product solving and how is that benefiting you?
Semantic search for now.
Production-ready vector database to get you started quickly
What do you like best about the product?
- Good documentation and usage examples
- Easy-to-use Python SDK
- Production-ready with low latency at our scale (10-20M vectors)
- Good integration with the AI/LLM ecosystem
- Easy-to-use Python SDK
- Production-ready with low latency at our scale (10-20M vectors)
- Good integration with the AI/LLM ecosystem
What do you dislike about the product?
- did not find an easy way to export all vectors that we needed for data science/cleaning
- will get expensive when hosting 100s of millions of vectors
- will get expensive when hosting 100s of millions of vectors
What problems is the product solving and how is that benefiting you?
We use Pinecone as a vector database for retrieval augmented generation using LLMs.
Solid Hosted Vector DB
What do you like best about the product?
Ease of deployment! It takes just a few minutes to get an index set up and deployed.
What do you dislike about the product?
The web-based API console could be improved, for example for experiments with metric (cosine vs dotproduct vs euclidean).
What problems is the product solving and how is that benefiting you?
Storing embeddings for RAG.
showing 21 - 24