Published: 2026-08-02
Analysis & perspective

Vector databases and RAG: how agents keep long-term memory

Name: Vector databases and RAG: how agents keep long-term memory
Uploaded: 2026-08-02
Description: A short explainer from Tech With Tim on why agents use a vector database and retrieval-augmented generation for long-term memory, and how that differs from the conversation history in the context window.

Chapters / key moments (click to jump — plays here on the page)

A 36-second explainer on the standard pattern for agent long-term memory: convert your data into embeddings — numerical representations of meaning — store them in a vector database, and give the agent a search tool so it retrieves only the relevant pieces for the task at hand. Tech With Tim names this as RAG (retrieval-augmented generation) and frames it as the answer to a specific problem — the context window is already full of the current conversation, so documents, Q&As and long-term facts have to live somewhere the agent can query on demand.

Watch on YouTube → · Related: why the conversation is re-sent every turn · ← Back to News

Vector databases and RAG: how agents keep long-term memory

More OpenClaw & Claude Code news

Go deeper: OpenClaw guides