What it does
This n8n template watches a Google Drive folder for new PDF, TXT and JSON files. It splits each file, creates embeddings with OpenAI's text-embedding-3-small model, and stores them in Postgres with the PGVector extension. Processed files move to a separate vectorized folder so they are not handled twice. It can run manually or on a daily schedule, 3 AM by default.
Use cases
- 01Preparing internal documents for an AI assistant
- 02Searching private files by meaning
- 03Indexing new files in a folder without repeats
Connects
Google Drive · Embeddings OpenAI · Recursive Character Text Splitter · Default Data Loader · Postgres PGVector Store