Team SQL or Team Vector DB for Agent Workflows? (We ditched vectors 👀)

Hi, Product Hunt🙋，

🇬🇧Our team in the UK is building an academic agent (PagePeek) that relies on 15+ specialized agents using the OpenAI Agent SDK. To solve context loss, we ditched Vector Stores for a PostgreSQL-backed system.😊

This allows strict handoffs between agents (e.g., Research -> Drafting) and supports our XML citation parsing for anti-hallucination.

Question: For complex agentic workflows, are you Team SQL or Team Vector DB? Let's discuss! 👇