How to Give Claude Persistent Memory (No Coding Required)
If you use Claude regularly, you've hit this wall: you spend 10 minutes giving Claude context about your project, your preferences, your background — and the next day you open a new conversation and have to explain everything again.
This isn't a bug. It's how Claude works by design. Every conversation starts completely fresh.
But there's a fix.
Why Claude Forgets Everything
Claude doesn't store anything between conversations. Each new chat is a blank slate. This protects your privacy but creates real friction for anyone using Claude for ongoing work — projects, learning, writing, coding, or anything that spans multiple sessions.
The Old Workarounds (And Why They Don't Work)
Most people try one of these:
**Copy-pasting context at the start of every chat** — works, but eats tokens and takes time. If your context is long, it also pushes Claude's context window toward its limit faster.
**Claude Projects with uploaded files** — better, but still locked to Claude. If you use ChatGPT or Gemini for some tasks, you're back to square one.
**Manually keeping notes in Notion or Google Docs** — you still have to paste them in manually every time.
None of these actually solve the problem. They just move it around.
The Real Fix: An External Memory Layer
The right solution is to store your memories outside of Claude — in a dedicated memory vault that Claude can search every time you start a conversation.
This is exactly what Lumi does.
How Lumi Works With Claude
Lumi connects to Claude via MCP (Model Context Protocol) — the same standard that lets Claude use external tools. Once connected, Claude can automatically save and retrieve memories from your Lumi vault.
Here's what that looks like in practice:
**Saving a memory:** You tell Claude "save this to my Lumi vault" and it does it immediately. Or you tell Claude "from now on, save everything important from our conversations automatically" — and it does that too, without you asking every time.
**Retrieving memories:** The next time you open Claude — even days or weeks later, in a completely new conversation — Claude searches your Lumi vault for relevant context and uses it in its response. You don't have to paste anything. Claude just knows.
Setting It Up (60 Seconds)
No coding. No config files. No terminal.
1. Go to **claude.ai** → Settings → Connectors
2. Click **"Add custom connector"**
3. Paste: `https://www.llmmemory.ai/api/mcp`
4. Click Add → Authorize when redirected to Lumi
5. Done
That's it. Claude now has access to your Lumi memory vault in every conversation.
What to Save First
Once connected, try these to see it working:
Open a brand new Claude conversation the next day and ask: **"What do you know about me?"**
Claude will retrieve your memories and respond with everything you've saved — without you saying anything else.
Works Across Claude, ChatGPT, and Gemini
The biggest advantage of Lumi over Claude's built-in memory: the same memories work across multiple AI tools.
Save something in Claude. Open ChatGPT. It knows it too. Switch to Gemini CLI. Same memories.
One memory library, every AI tool.
Free to Start
Lumi has a free tier — 300 memories and 2 vaults. No credit card required.