I want to set up a system where each file corresponds to its own vector database because combining different files requires rebuilding the vector database. This approach is intended to reduce the cost of vectorization using OpenAI’s embedding model. However, the issue I’m facing is that the OpenAI API for assistants only supports a single vector database. How can I achieve a system that allows retrieval across multiple vector databases?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| One Assistant Multiple Vector Stores? | 1 | 1330 | June 20, 2024 | |
| Assistants API with local vector store | 2 | 698 | June 24, 2024 | |
| Creating my own vector store for assistant api | 2 | 1111 | July 24, 2024 | |
| Specific files per thread for assistants api | 3 | 711 | July 18, 2024 | |
| Assistant API with external Vector stores | 3 | 524 | August 2, 2024 |