RAG Document Ingestion (Chunk → Embed → Store)
Turn any document into searchable knowledge: receive text via webhook, split it into overlapping chunks, generate OpenAI embeddings, and store each chunk in a Supabase (pgvector) table. Pairs with the RAG Chatbot template.
Steps in this flow
- Webhook: Document In
- Split into Chunks
- For Each Chunk
- OpenAI: Embed Chunk
- Supabase: Store Chunk