Dhruv Kumar Jha
Dhruv Kumar Jha8mo ago

Querying embeddings

For querying embeddings, do we have to use the node runtime? I was looking at the convex Langchain integration and it's mentioned over there but not in convex website docs https://docs.convex.dev/vector-search
Vector Search | Convex Developer Hub
Run vector search queries on embeddings
2 Replies
Michal Srb
Michal Srb8mo ago
I'm not aware of any need for Node runtime around embeddings. You should be able to call OpenAI and other LLMs from normal actions, and read and write to the DB from queries/mutations.
ian
ian8mo ago
The Langchain npm module is huge and has a ton of gnarly dependencies, which I believe was the reason for putting it in the node runtime. also recommended to put it in the external dependencies list if you use it

Did you find this page helpful?