How to solve n8n Supabase Vector Dimensions Embedding error
If everything works fine except the Supabase part where the vector data is stored, then there might be a vector dimensionality error.
In my case, I was using a text embedding model by Google which generated 768-dimensional vectors, but the default table in Supabase was expecting 1536-dimensional vectors (default for OpenAI/GPT embeddings). This mismatch was causing the issue.
You can solve this in 2 ways:
1. Alter or change the SQL table in Supabase using SQL commands in the editor to match the correct dimensionality.
2. Delete the existing table along with its dependencies and recreate a new one with the correct vector dimensionality.
#n8nworkflow
#automation #solution #n8n