anything-llm/server/storage/vector-cache/VECTOR_CACHE.md

6 lines
487 B
Markdown
Raw Normal View History

2023-06-04 04:28:07 +02:00
### What is this folder?
`vector-cache` is a running storage of JSON documents that you have already run embeddings on. This allows you to use the same large documents for multiple workspaces without having to pay to re-embed them each time you want to reference them across workspaces.
This also allows you to reset entire workspaces back to their original state without having to pay for the embeddings again. Saving you tons of money for large documents that take a while to embed.