Early access · one memory, every AI tool

Stop re-explaining yourself to AI.

Every ChatGPT, Claude, Gemini, Cursor, and Claude Code session becomes a dot in your memory graph. Open any tool tomorrow and the right context is already there. One URL connects them all.

Works with ChatGPT Claude Gemini Cursor Claude Code · one memory across them all

Your context is scattered across vendors.

Every new chat starts from zero. You've explained your stack to ChatGPT, your architecture to Cursor, your plans to Gemini. Tomorrow you'll explain them all again. Each tool keeps its own memory, locked inside its own walls, and every re-explanation burns tokens you already paid for.

MemoryOS sits underneath all of them. One graph holds every session and serves it wherever you work next, from the chat app on your phone to the IDE at your desk. Connect it once. It follows you everywhere.


How it works

01

Capture

When a session concludes, your assistant saves it as a dot: a compact summary of what was decided, with its source and tags. Add one standing instruction and this happens on its own.

02

Connect

New dots link themselves to related ones through shared topics. On Max, an AI agent adds typed links with reasons. No manual filing. The graph assembles itself as you work.

03

Recall

Start a new session and ask. MemoryOS serves the matching dots and their connections as a compact context block. Need one specific memory? Copy its reference from your graph and paste it in any connected tool.

What it does

A graph, not a pile

Sessions connect through shared topics into a navigable web, a graph view for your AI history. Hubs emerge around the things you actually work on.

Cross-tool by design

One memory across ChatGPT, Claude, Gemini, and your IDE. Built-in memories are vendor silos. This one travels with you, phone to editor.

Pick up where you left off

Research something in ChatGPT on your phone, then continue in Cursor at your desk. Every dot has a permanent reference, so the exact context follows you, ready to build on.

Self-assembling

Dots auto-link on save when they genuinely share content, and the Max agent labels each link with why it exists. Connected, never a hairball.

Recall that serves, not dumps

Hybrid retrieval: full-text search fused with semantic similarity, expanded along the graph, rendered into a token-budgeted block a model can actually use.

Built on MCP, the open standard

The connector protocol ChatGPT, Claude, Cursor, and VS Code speak natively. One OAuth-secured URL connects them all, plus a personal API when you want to script your memory.

Pricing

Basic

$0

Free. Make the habit stick.

  • 50 dots of memory
  • Chat apps: ChatGPT, Claude, Gemini
  • Short served context (3–4 lines a dot)
Start free

Max

$19/mo

or $189/yr. For heavy daily use.

  • Everything in Premium
  • ✦ The MemoryOS Agent: chat to organize your graph into packs
  • AI links between your dots, each with a reason
  • Huge context + personal API token
Get Max

Questions

How do I connect a tool?

With one URL. Paste it as a custom connector (MCP) in ChatGPT, Claude, Cursor, Claude Code, Windsurf, or VS Code, approve access, and you're done. The docs have exact steps for every tool.

What actually gets stored?

Only dots: short, self-contained summaries your assistant saves when a session concludes. Never full transcripts unless you paste one in yourself. Every dot is visible in your graph, so you always know what your memory holds.

Does it work on my phone?

Yes. Set the connector up once on desktop web; after that the ChatGPT and Claude mobile apps carry your memory too. Research on the couch, build at the desk.

Can I move one specific memory between tools?

Yes. Every dot has a permanent short reference. Copy it from your graph, paste it in any connected tool, and that dot is served right there with its connected context. It's how a phone research session becomes a desk coding session.

Is my memory private?

Each account's dots live in their own isolated database, never mixed with anyone else's. Connections use OAuth 2.1 with PKCE, traffic is HTTPS, and your dots are served only to tools you explicitly approved.

What happens when I hit Basic's 50 dots?

Nothing is deleted and recall keeps working. You just can't save new dots until you upgrade. Caps apply to saving, never to remembering.

Early access · one memory, every tool

Your AI should remember you.

One connected memory for every assistant and IDE, served where you work.