I've used Pinecone to streamline token generation for my chatbot's functionality. Specifically, I used it for the OpenNeeam Building.
Reviews from AWS customer
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
External reviews are not included in the AWS star rating for the product.
ideal for machine learning, AI applications and similarity search
Also it's use case is little complex with lack of ecosystem integration.
God of creating embeddings
Pinecone assistant beta user
I love to see pinecone assistant in deployable version because it is powerful yet it is in the beta version only for testing not for production
Good support from team.
Solid option for vector DB
Pinecone: The Backbone of Efficient Vector Search and Retrieval
Using Pinecone on production - 1 year later
- High performance (upsert and search in the ms)
- Simple integration via API and deployment and now after their recent release of serverless indexes it's very simple to maintain and scale (it's autoscaling).
- Low price (relative to the number of vectors) and free limited indexes. Free indexes are great to run development environment data. For a while it was impossible to upgrade a free index to a paying one, but this is now addressed.
- Incredible support (we had an issue and was not expecting getting this quality of support without paying the usual business support fees of an AWS for example)
- The ability to assign metadata is very useful (we still maintain a traditional db to keep track of the vectors)
- The single stage query vector/metadata is very useful and saves the headache of over-querying
- One feature we have meant to use is the use of sparse vectors in combination with the dense vectors. So, can't really comment yet
- The documentation using metadata and single stage queries is a bit light
- They have a smart bot to help answer support questions. On the great side, it seems they use their own technology for RAG type of application, but on the other it often misses the mark. ChatGPT or Perplexity are surprisingly more effective.
- There has been a few down times, but they are very communicative about them and maintain a server health page for each endpoint. It's usually related to a specific infrastructure (AWS or GCP) they run on.
- They have been growing and improving the technology, and like with other player, sometimes to update their python library or the way to reference to the indexes. But each time it's been toward simplification, and I suspect it will stabilize.
A great serverless DBaaS for vectors
Provides a private local host feature and is easy for new users to learn
What is our primary use case?
What is most valuable?
The best thing about Pinecone is its private local host feature. It displays all the maintenance parameters and lets us view the data sent to the database. We can also see the status of the CD and which application it corresponds to.
What needs improvement?
I want to suggest that Pinecone requires a login and API key, but I would prefer not to have a login system and to use the environment directly.
For how long have I used the solution?
I have used Pinecone for the past three months.
Which solution did I use previously and why did I switch?
Before Pinecone, I used Corner DB.
How was the initial setup?
The installation of Pinecone was straightforward.
What's my experience with pricing, setup cost, and licensing?
I think Pinecone is cheaper to use than other options I've explored. However, I also remember that they offer a paid version.
Which other solutions did I evaluate?
I decided to use Pinecone after researching and finding it the best option for our project.
What other advice do I have?
Pinecone is easy for new users to learn, and I would rate it around eight out of ten. This is because other databases do not have a login system and are not as user-friendly.
Which deployment model are you using for this solution?
It is very flexible, allowing us to input any kind of data dimensions into the platform
What is our primary use case?
I used Pinecone in collaboration with an Azure database. At that time, I needed to create a chatbot that could pull data from public media in specific fields. I used Pinecone to embed the publications, and after submitting the data, it was pushed into our data pipeline.
What is most valuable?
The most valuable feature of Pinecone is its managed service aspect. There are many vector databases available, but Pinecone stands out in the market. It is very flexible, allowing us to input any kind of data dimensions into the platform. This makes it easy to use for both technical and non-technical users.
What needs improvement?
Pinecone can be made more budget-friendly.
For how long have I used the solution?
I have been using Pinecone for the past year and a half.
What do I think about the stability of the solution?
Pinecone is a stable product. Despite few errors, it's easy to use, especially when searching with endpoints. Compared to other databases, Pinecone is quite user-friendly.
What do I think about the scalability of the solution?
Pinecone is a scalable product. We can easily add users and workload without any issues.
Which solution did I use previously and why did I switch?
How was the initial setup?
The installation, setup, and deployment of Pinecone is straightforward. We need to take a subscription from Pinecone and configure the endpoints into our applications. Before configuration, we need to install Pinecone libraries on the dev side. We put the tokens at the endpoints and connect Pinecone to our applications. After that, we push our metadata into the Pinecone endpoint database. Once the data is pushed, we can search the data we've entered. Pinecone supports various functions based on similarity and allows us to specify how many results we want, like the top five or top two results.
What's my experience with pricing, setup cost, and licensing?
Pinecone is not cheap; it's actually quite expensive. We find that using Pinecone can raise our budget significantly. On the other hand, using open-source options is more budget-friendly.
Which other solutions did I evaluate?
We chose Pinecone because other vector databases, like ProMID or Azure, don't have UI-rich components or tools. Pinecone offers a better UI and allows us to create any kind of application and handle a large amount of data easily. It is a managed service, making it more convenient for us.
What other advice do I have?
As per my advice, assess your data requirements. If you're working with PDF files and do not have much data, you could use other databases because they are similar to Pinecone. However, if you have a huge amount of data, I would suggest using Pinecone as it handles large datasets more efficiently. Pinecone offers a rich UI and managed services, making it easy to use and visualize data, which is a big advantage. However, if the client has a limited budget, I would recommend open-source models and databases instead.
I would rate Pinecone an eight out of ten because of its functionality and ease of use despite the cost.
Which deployment model are you using for this solution?
Helps retrieve data, relatively cheaper, and provides useful documentation
What is our primary use case?
Pinecone is a vector database. We use it to retrieve data using semantic search. We use vector DB only for chatbots and AI applications. Currently, I am using the tool to make a chatbot.
What is most valuable?
The semantic search capability is very good. We store data and embed numeric values. If I want to search for something, I get the right data 90% of the time.
What needs improvement?
Suppose I want to delete a vector from Pinecone or a multi-vector from a single document. Pinecone does not provide feedback on whether a document is deleted or not. In SQL and NoSQL databases, if we delete something, we get a response that it is deleted. The tool does not confirm whether a file is deleted or not. I have raised the issue with support.
If we have 10,000 vectors in our index and do not use a metadata tag, it will take one to three seconds to complete a search. When I try to search using a metadata tag, the speed is still the same. The search speed must be much faster because I specify which vectors I need the data from.
For how long have I used the solution?
I have been using Pinecone for almost one year.
What do I think about the stability of the solution?
I face some breakdowns. However, it happens rarely. Sometimes, the server crashes when we retrieve data from it.
What do I think about the scalability of the solution?
We have a SaaS project, and Pinecone is a database for that project. All the developers who work on the project use the solution. Currently, six to seven of us use the solution. We recently moved to serverless DB. It is easy to create metadata fields. If we have a certain template for our database, we can change the database very easily. It will not show any errors. We just have to put an extra key in the metadata fields.
How are customer service and support?
I was unable to delete the data using IDs and metadata. I raised a query for it. I got the response in less than 24 hours, and it was resolved. The support team is very good. They provide quick responses.
How was the initial setup?
The solution is deployed in the cloud. The tool is very easy to install. There are commands to install the tool. The product is very easy to install and integrate on our machine.
What's my experience with pricing, setup cost, and licensing?
Initially, the product was expensive. My company used to pay $70 per index. Now, we can pay according to our needs. It is a pay-as-you-go model. For the same use case, we are currently paying $4. The solution is relatively cheaper than other vector DBs in the market. It is worth the money.
Which other solutions did I evaluate?
We also use Weaviate for some projects. It is also a vector DB. We also use an SQL database called PlanetScale. Before installing Pinecone, we compared its performance against vector databases like Weaviate and ChromaDB. Pinecone and Weaviate emerged as the top choices.
What other advice do I have?
Pinecone and Weaviate are both good choices. If we want to use the solution, we must know how a vector DB works theoretically. Then, we will be able to work with it easily. If we do not know how vector DBs work, we must refer to the documents to insert and get data. Having a basic understanding of vector DBs is helpful. If a beginner goes through the documents, it is very easy to use the product.
Overall, I rate the product an eight out of ten.