Basic · about 2 minutesBest place to start
Move what ChatGPT remembers now.
Copy one prompt into ChatGPT, review the profile it writes, and every AI you use knows you from that moment on.
ChatGPT → Claude in 2 minutes
Never explain yourself twice.
Move what ChatGPT remembers about you to Claude, then keep one memory across every AI you use.
14 days free. $20/month after. Cancel anytime.
Your memory
Follows you everywhere
Yours
What it already knows about youToday
How you like answers written
What you’re working on right now
People, projects, and context you’ve shared
The last thing you asked it to remember
Used from
ChatGPTClaudeCoworkClaude CodeCodexOpenClaw
Next answer starts with
context
not a blank page
ChatGPT remembers your chats. MemoryRouter remembers you, across every AI you use.
The magic moment
Give ChatGPT's memories to Claude in 2 minutes.
You spent months teaching ChatGPT who you are. Take all of it with you. Claude knows you the moment you say hello.
01
We give you a prompt.
Copy it into ChatGPT. Takes about a minute.
02
ChatGPT hands over everything it knows about you.
You watch it happen and pick exactly what saves.
03
Claude picks up mid-sentence.
No blank page. No re-explaining who you are. Ever again.
Two minutes. Then it works in every AI you touch.
Choose your transfer
Start simple. Go deeper when you want to.
Basic moves everything ChatGPT remembers about you. Advanced digs the exact details out of years of old conversations.
Advanced · official exportFor years of old chats
Recover exact details from old conversations.
Drop in your official ChatGPT export and the Advanced importer pulls the exact details out of every old chat. Preview, approve, done. Years of context, yours again.
Works with your tools
Your AI should know you by now.
One memory across ChatGPT, Claude, Claude Code, Codex, OpenClaw, DeepSeek Harness, and whatever ships next. Pick a tool to see the two-minute setup.
Connector address
https://mcp.memoryrouter.ai/mcp
Paste it in Settings, click Connect, sign in. Done.
Connector↗
ChatGPT
Add MemoryRouter as a connector and ChatGPT keeps writing to a memory that never fills up and never stays trapped in one app.
Paste URL, connect, sign in
Connector↗
Claude and Cowork
Open Connectors, add the MemoryRouter address, and Claude knows everything ChatGPT knew about you from the first message.
Customize, Connectors, Connect
Coding memory↗
Claude Code
Hooks recall your project context before every prompt and save every finished turn. Compaction and new sessions stop costing you anything.
npx memoryrouter-claude init
Coding memory↗
Codex
One command connects Codex. Every new session opens already knowing your decisions, your stack, and where you left off.
codex mcp add memoryrouter
Agent memory↗
OpenClaw
Two commands give your OpenClaw agent a memory that survives compaction, restarts, and model switches.
openclaw plugins install
Agent memory↗
DeepSeek Harness
One command gives dsh a memory that outlives every session. Every channel, one memory. Clean user and assistant text, zero tool noise.
dsh plugin add dsh-memoryrouter
Connector↗
Any MCP client
Cursor, Windsurf, Gemini CLI, and every other MCP-compatible tool plug into the same memory with one address.
https://mcp.memoryrouter.ai/mcp
Terminal↗
CLI
Search, store, and manage your memory from the terminal. Pipe context into any script or workflow you already run.
npx memoryrouter
Never explain again
Say less. Do more.
This is what changes the day your memory travels with you instead of dying in one chat window.
Every new chat starts from zero.
Every new chat starts with everything. Say less, get more.
You keep re-pasting the same context, links, and preferences.
Your memory follows you. Paste once, never again.
AI amnesia resets the moment you close the tab.
Close the tab. Open a different AI. Pick up exactly where you left off.
Trying a new model means losing everything you taught the old one.
Switch models whenever you want. Your memory is yours, not the vendor's.
Outcomes
What you get back
Every new chat and every new model starts warm. That is hours back every week.
Never re-explain
Your background, preferences, and project state are already loaded before you type a word.
Zero copy-paste
Your context moves with you. Nothing lives in old chats, notes apps, or a graveyard of browser tabs.
Cheaper every month
The right context is already there, so you stop paying tokens to restate what your AI should know.
Faster from the first message
Ask once, reuse everywhere. Drafts, research, and answers start closer to done.
Start here
One memory, every AI you use.
Connect the AI you use most today. Add the rest whenever you like. The memory is the same everywhere, and it is yours.
Starter
$20/mo
200M memory tokens included every month. Every AI you use. Every model.
14 days free. Cancel anytime. No inference markup, ever.
More ways to use MemoryRouter
Here for your team, or building your own?
For teams and companies
Your whole team shares one memory.
Every teammate's AI knows the company on day one. Shared context, access controls, and a rollout that takes an afternoon.
See MemoryRouter for teamsFor developers
Give the agents you build a memory that lasts.
API, MCP server, CLI, and framework guides. Persistent memory for every user of your product, wired in an afternoon.
Read the developer docs