Help! N8N Pinecode Vector Store Node does not support the new Pinecone Llama-text-embed-v2 embedding.
Pinecone has replaced the OpenAI text-embedding-3-small embedding by the Llama-text-embed-v2 embedding to create new indices. But for the N8N Pinecode Vector Store Node this new embedding is not yet supported.
Existing workflows with Pinecone Vector Store Nodes with OpenAI text-embedding-3-small embeddings are still running but you cannot create new ones.
The dimensions of the OpenAI and the Llama embeddings are different.
I use the Free Plan of Pinecone but I guess this also is true for the other plans.
Does anyone have more information about this topic?
Hi Sandy, this is a misunderstanding. You can still create an OpenAI text-embedding-3-small embedding for a Pinecode Vector Store, but you cannot create a text-embedding-3-small based index in Pinecone anymore. On the other hand in N8N there is no embedding available for the new Llama-text-embed-v2 model.
As a result you cannot create a N8N Vector Store Node and choose an index that will be compatible with your embedding.
for me it is NOT possible to scroll to the right, too - but..
i have to choose between “Integrated Embedding” and “External Embedding”. Choose second one (“External Embedding”) and then “Autofill values from model”. Here you get a list with models containing “text-embedding-3-small” - that’s it.
