How to Give Claude
Perfect Memory

Claude’s memory is pretty bad out of the box. It saves random facts, forgets important ones, and loses everything the second you start a new chat. Most people have been living with this for months without knowing there’s a better way.

Here are three layers of memory systems that change everything:

Layer 1 — Basic Memory (Beginner) — 4 quick wins, takes minutes
Layer 2 — Context File System (Intermediate) — ~60 min setup, changes how Claude operates entirely
Layer 3 — AI Second Brain (Advanced) — turns Claude into a self-evolving system trained on all your data

Start wherever you are. Each layer builds on the last.

This is where everyone should start. Four quick fixes that take minutes and immediately improve every conversation.

1. Clean Up Your Memory Page

Go to Settings → Memory right now. This is the most overlooked page in all of Claude. You’ll see everything Claude has stored about you — preferences, facts, habits, working styles — accumulated across every conversation you’ve ever had. Left unmanaged, it fills up with garbage. Read through everything. Delete anything outdated, inaccurate, or irrelevant. Then manually add the context you actually want Claude to carry permanently.

2. Fill In Your Project Instructions

If you use Claude Projects (you should), fill in your Project Instructions field for every project. This tells Claude the context for that specific workspace — your role, your rules, your audience, your files. Create projects for your most-used workflows, then give each one detailed instructions so Claude never starts from scratch.

3. Tell Claude Directly

The simplest hack on this list. Mid-conversation, just tell Claude what to remember:

• “Remember that I never want bullet points in emails”
• “Remember that my role is [x] at [company]”
• “Update your memory: I prefer responses under 400 words”
• “Forget that I mentioned [x]”

Claude stores these immediately. You can verify by checking your Memory page after.

4. Import From ChatGPT (or Other LLMs)

If you’ve built up context in ChatGPT, you don’t have to start from scratch. Two options:

Option A: Tell ChatGPT you’re switching platforms and ask it to generate a memory export: “I’m switching to Claude. Give me a complete summary of everything you know about me, my preferences, my work, and my communication style.” Copy that output and paste it into Claude.

Option B: Use Claude’s built-in import tool. Go to Settings → Memory and look for Import/Export. Claude can import full data from other LLMs directly.

Layer 1 fixes the basic memory problems. Layer 2 builds something more powerful: a file-based memory system that lives on your computer and loads automatically into Cowork and Claude Code.

Instead of relying on Claude’s built-in memory (which is unreliable), you store all your context in .md files on your desktop that Claude reads every time you work. You can also attach these files to any LLM or AI agent system.

The 4 Files You Need

1. Instructions.md — Tells Claude all your rules and instructions. Who you are, what you do, how you want things done, what good output looks like. Important: include the line “Update Memory.md with my preferences over time” — this is how Claude creates a running memory log automatically.

2. Memory.md — This is the “brain” of Claude. It starts mostly empty and gets updated over time as you work. Whenever you say something like “stop using em dashes” or “I prefer shorter emails,” Claude goes into this file and updates it. Sections: Preferences, Corrections, Patterns, Decisions.

3. Context.md — The specific context for whatever project you’re working on. Your business, your audience, your goals, your active priorities. You can create a general “business context” file or separate ones per project.

4. Archive Copies — Claude will update your memory files as you work. Occasionally it overwrites something incorrectly. Without a backup, that context is gone. Once a week, copy your entire master folder into a separate archive folder that Claude can’t access. Label it with the date. If anything breaks, restore from the archive.

You can build all 4 files yourself, or just paste this prompt into Cowork and Claude will build them for you:

This is the deepest level and it’s not for everyone. It requires setup and ongoing maintenance. But if you build it, Claude becomes a self-evolving second brain trained on everything you think, read, write, and work on.

There are two options depending on how you work.

Option 1: Claude + Notion (Easier)

Connecting Claude to Notion is the highest-leverage thing you can do in 5 minutes.

Step 1: Go to Claude → Settings → Connectors and enable the Notion connector.

Step 2: Once connected, Claude can read your entire Notion workspace directly inside any chat. All your tasks, CRMs, notes, tables — instantly accessible and editable.

Step 3: Create a new Notion database called “Memory Database” where you store all your AI preferences, rules, and important context.

Step 4: As you work with Claude, tell it: “Send this to my Notion Memory Database.” Over time, this becomes a searchable, visual knowledge base of everything Claude knows about you.

Option 2: Claude + Obsidian (Most Powerful)

Obsidian stores everything as plain Markdown files on your computer. This makes it the most powerful option for building a second brain with Claude — because Claude can read, write, search, and evolve your entire knowledge base over time.

This is one system. The Weekend Bootcamp walks you through setting up Claude as a complete operating system for your specific job — memory, Projects, Skills, connectors, scheduled tasks — all wired together. One weekend.