Uploaded May 2026 | Updated September 2026, 2 weeks ago

---

Last time, we built a chatbot that could answer questions using Oracle AI Vector Search.

But it still had one major problem:

It forgot everything between requests.

In this video, we fix that by adding persistent chat memory with Spring AI and Oracle Database.

The app now:

* remembers conversations
* keeps memory scoped per user/session
* persists context across restarts
* combines RAG + conversational memory

We’ll look at:

* JDBC-backed chat memory
* `MessageChatMemoryAdvisor`
* conversation IDs
* Oracle-backed persistence
* why memory is different from retrieval

One of the key ideas in this episode:

RAG helps the model know things.
Memory helps the model remember what *you* said.

Repo:
[github.com/markxnelson/shopassist](https://github.com/markxnelson/shopassist)

#springboot #springai #oracle #ai #rag #java