Give Claude Infinite Memory With NotebookLM

What You'll Have

By the end of this guide, you'll have a simple memory loop for Claude.

• NotebookLM stores your long-term project knowledge.

• Claude receives only the relevant context for the current task.

• Each Claude session ends with a clean memory update.

• Future sessions start faster because you do not have to re-explain the same project every time.

Core idea: Claude is the worker. NotebookLM is the memory bank. Store source material in NotebookLM, retrieve the relevant memory, then paste that focused context into Claude.

Who This Is For

• You use Claude, Claude Code, or Claude CoWork for ongoing work.

• You keep repeating the same brand, project, or client context.

• Your Claude chats get messy because every session starts from scratch.

• You want a lightweight memory system without building a database or RAG setup.

• You want past decisions, research, and rules to stay available across sessions.

What You'll Need

• A Claude account.

• A NotebookLM account.

• One dedicated NotebookLM notebook for the project.

• Project sources such as docs, transcripts, briefs, URLs, notes, SOPs, videos, and prior decisions.

• A simple habit of saving session updates back into NotebookLM.

Going Further

• Use one NotebookLM notebook per project or client.

• Create a weekly memory update summary.

• Keep one source called Project Source of Truth.

• Add only reviewed and useful material.

• Ask NotebookLM for focused retrieval before every major Claude session.

• Ask Claude for a memory update before ending every session.