Claude Desktop / claude.ai
One-click install. Sign in with Google when prompted — no API key to copy.
View setup →Portable memory for ChatGPT, Claude, Cursor and Slack
One memory. Every AI. Nothing to install.
Memxus works across Claude, ChatGPT, Cursor and Slack — no extension, no local setup, no repeated context. Save once. Remember everywhere.
Start free. No credit card required.
No installation. No browser extension. No weird prompts. Full control over what gets saved.
Built for AI power users, developers and teams who work across multiple AI tools.
ChatGPT
User: Remember that this project uses Next.js, Supabase and MCP.
Memxus: Saved.
Claude
Claude: I found your project context: Next.js, Supabase and MCP.
Cursor
Cursor: Using your saved project memory to suggest the right implementation.
One shared memory layer
Where Memxus sits in your stack
Not another chatbot — the shared layer between your AI tools and your data.
Persistent memory
Cloud memory
Save from anywhere
Claude
ChatGPT
Slack
Memxus
Portable memory layer
MCPAPIOAuth
Recall anywhere
Claude
Cursor
Gemini
remember·recall
Your memories
Persistent · Private · Editable
Dashboard · Notebook
View, edit, delete anytime
Remember (save)
Recall (use elsewhere)
Your AI tools don’t share context
You explain something to ChatGPT. Claude doesn’t know it. Cursor doesn’t know it. Slack doesn’t know it. Every tool starts from zero.
Repeated context
You keep explaining the same projects, preferences and decisions again and again.
Fragmented AI workflows
Each AI assistant has its own memory, its own context and its own blind spots.
Lost decisions
Important project details disappear across chats, tools and teams.
Use fewer tokens. Keep more budget.
Stop paying to re-send the same context in every chat.
70%
less repeated context in a typical session
2,400
tokens saved per session (median estimate)
$8–25
estimated monthly API savings for daily users
Without Memxus
6,000 tokens
Full project brief pasted each time
With Memxus
1,200 tokens
Targeted recall via get_context
Pro plan from $12/mo — often less than what you save in tokens alone.
Estimates based on typical context sizes and public list prices (e.g. Claude/GPT input ~$2–5 per 1M tokens). Your savings depend on model, session length, and how much you repeat context.
Memxus gives every AI the same context
A portable memory layer that works across your AI tools, apps and workflows.
1
Connect
Connect ChatGPT, Claude, Cursor, Slack or your app through MCP, API or native integrations.
2
Save
Save facts, preferences, project context and decisions as structured memory.
3
Recall
Memxus retrieves the right context when you need it, wherever you are working.
Connect your AI
Pick the client you use every day. Each path takes a few minutes.
Cursor
HTTP MCP with your aimem_* API key from the dashboard. Six tools in Settings → MCP.
View setup →VS Code
Copilot MCP via mcp.json — same API key, different JSON format than Cursor.
View setup →ChatGPT
Via Custom GPT Actions (not MCP — ChatGPT doesn't support it yet)
Does not use MCP. Wire a Custom GPT with Actions to the REST API and your aimem_* key.
View setup →Built for real AI workflows
Automatic recall
Retrieve relevant memory without rewriting long prompts every time.
Editable memory notebook
See, edit, delete and organize everything Memxus knows.
Temporary mode
Choose when something should not be saved.
MCP-ready
Use Memxus with MCP-compatible tools, agents and AI clients.
Team memory
Share project context, decisions and knowledge across your team.
API-first
Add persistent memory to your own apps, agents and workflows.
Use it wherever context matters
For developers
Keep project architecture, stack decisions, coding preferences and implementation context available in Cursor, Claude and ChatGPT.
For founders
Preserve product strategy, customer insights, launch plans and positioning across every AI session.
For teams
Create shared memory for projects, workflows, decisions and recurring knowledge.
For AI power users
Stop rebuilding context every time you switch between models and tools.
You control your memory
Memxus is designed around transparency, editing and user control.
- View everything that has been saved
- Edit or delete any memory
- Use temporary mode when you don’t want persistence
- Control connected tools and integrations
- Export your memory when needed
Memory Notebook
Project: Memxus
Stack: Next.js, Supabase, MCP
Preference: Short, practical answers
Status: Editable
EditDeleteDisable
See how portable memory works
Save once in ChatGPT. Ask Claude or Cursor — they recall the same memory.
ChatGPT
Memxus
- 1.Save once in ChatGPT
- 2.Ask Claude to recall
- 3.Same memory returned
- 4.Ask Cursor to recall
- 5.Same memory in the IDE
- 6.Manage in Notebook
FAQ
Stop starting from zero
Connect your first AI and create a memory layer that follows your workflow.