Pricing for Vector storage and bandwidth
Pricing for vector storage and bandwidth is not clear to me, It would be helpful if anyone can provide any assistance.
NOTE: this is just an example to help me understand pricing better, you're welcome to correct me.
Also, please do not take into consideration what convex includes in their pricing tier, consider these are additional usage
Consider this example:
I have a document with 1000 pages and each page has 500 words.
Vector Dimension: 1536
Total words = 1000 * 500 => 500000 words in total
Now I want to store the embeddings in vector database
I convert each page into 10 chunks => 10 chunks * 1000 pages => 10,000 vector db records (only embeddings and ids of related tables (3) fields, not storing texts here)
Since 1 page has 500 words, Let's assume each chunk contains 500/10 => 50 words => approx 70 tokens.
Question: How much would it cost me in vector storage for these 1000 pages?
---
Now let's start querying the vector db, I query vector database 1000 times with my embeddings, And I request 10 results per query.
Question: How much would it cost in bandwidth for 1000 queries with 10 result per query?
Also, I love the flexibility Convex offers and I want to use it but would it be ideal for SaaS RAG application?
If I have 1000 users uploading 1000 pages each and querying vector store 1000 times in a month - how much would it cost me?
NOTE: this is just an example to help me understand pricing better, you're welcome to correct me.
Also, please do not take into consideration what convex includes in their pricing tier, consider these are additional usage
Consider this example:
I have a document with 1000 pages and each page has 500 words.
Vector Dimension: 1536
Total words = 1000 * 500 => 500000 words in total
Now I want to store the embeddings in vector database
I convert each page into 10 chunks => 10 chunks * 1000 pages => 10,000 vector db records (only embeddings and ids of related tables (3) fields, not storing texts here)
Since 1 page has 500 words, Let's assume each chunk contains 500/10 => 50 words => approx 70 tokens.
Question: How much would it cost me in vector storage for these 1000 pages?
---
Now let's start querying the vector db, I query vector database 1000 times with my embeddings, And I request 10 results per query.
Question: How much would it cost in bandwidth for 1000 queries with 10 result per query?
Also, I love the flexibility Convex offers and I want to use it but would it be ideal for SaaS RAG application?
If I have 1000 users uploading 1000 pages each and querying vector store 1000 times in a month - how much would it cost me?
