Shared memory for your AI clients.
Notes, decisions, conventions, product docs — written once, then read and updated by whoever needs them next. One endpoint over MCP, or the API inside your own product.
Works with Claude, ChatGPT, Cursor, VS Code, Claude Code — and any MCP client.
Your AI has amnesia.
Every chat starts from zero. Details get lost, and context gets mixed up between chats.
Everything stays saved
Save a fact once and it stays — the next chat, the next tool, months later. Nothing to re-explain.
Data never gets mixed up
Each project keeps its own memory, so details from one client never leak into another.
Anything is one search away
Search by meaning, not keywords — across every memo at once, or inside one project and tag.
You’re locked in one AI tool.
What ChatGPT learns, Claude never sees. Switch the app and you start briefing from scratch again.
Works with every AI tool
One MCP server, all your clients — Claude, Cursor, ChatGPT and the rest read the same memory.
Same memory on every device
Save it on the laptop, use it on the phone. No exports, no copy-paste between apps.
Your knowledge stays yours
Change tools whenever you like — the memory belongs to you and simply follows along.
Everyone on the team knows something different.
Docs go stale, answers hide in someone’s chat, and every new person asks the same questions again.
One memory for the whole team
Invite teammates to a project and everyone — with their AI — reads and writes the very same memory.
Always the current version
Update a memo once and every teammate gets the new answer, not last month’s copy of the doc.
New people start up to speed
A new teammate’s AI reads the project memory from day one — nobody explains the basics again.
Working with a team? See Memocore for teams →
Set it up once. Never brief your AI again.
Memocore connects to the tools you already use and works quietly in the background.
Connect MCP
Add the Memocore MCP to your AI client — one click in Cursor or VS Code, a single line everywhere else.
Remember
Just ask your AI client to save what matters to Memocore — a decision, a preference, a project fact.
Recall
Ask your AI client to look it up in Memocore, and the right memo comes back — in any chat, in any tool.
Once you sign up, a step-by-step guide for connecting every AI client is waiting inside.
Anything else — just ask the chatbot in the bottom-right corner.
Give your product a memory over the API.
Connect your own app or chatbot to Memocore with a simple API. Instead of stuffing your whole knowledge base into every prompt, pull only what’s relevant.
A RAG layer without the stack
Search by meaning is built in. No vector store, no embeddings pipeline, no re-indexing — write memos, search them, ship.
Cut token costs
Fetch the few memos that matter, not the entire knowledge base. Smaller prompts, smaller bills.
One source of truth
Your product and your assistants read and write the same memory.
GET /api/v1/memos?query=refund+policyAuthorization: Bearer mc_live_•••{ "items": [ { "title": "Refund policy — EU" }, { "title": "Chargeback playbook" } ], "count": 2}
A memory layer built to be shared.
Private by default, collaborative when you want it, and portable across everything you use.
Works in every AI client
One MCP server, every assistant. No plugin per app, no copy-paste between tools — the memory is the same wherever you connect.
Private & team memory
Personal notes stay yours. Project memory is visible only to the teammates you invite.
Finds the right note
Meaning and keywords in one search, so the right memo surfaces before your AI answers.
Share with a link
Make a single memo — or a whole project — public. A clean page anyone can open, no login required.
Teams & projects
Invite members, set roles, and give a whole team one shared brain.
Files included
Attach docs and images to a memo, recalled right alongside the note.
Nothing changes silently
Every memo keeps an activity history — who edited it and when, so the whole team stays in the loop.
Connect your tools
Pull memory into Notion and Telegram, or wire up your own app over the REST API.
Start free. Scale when it sticks.
Every plan works across all your AI clients. Upgrade only when you outgrow the free tier.
Team
$15/seat/mo
- Unlimited memos
- Unlimited projects
- 5 GB storage per seat
- External integrations
- Share your memos
- Create public projects
Common questions
A persistent memory layer for AI. It stores your notes and knowledge, and lets every AI client you use read and write the same memory.
Any client that speaks the Model Context Protocol — Claude, ChatGPT, Cursor, VS Code, Claude Code and more.
Yes. Personal memos are private to you; project memory is shared only with the teammates you invite.
No. Connect one endpoint to your AI client — one click in Cursor or VS Code, a single line everywhere else.
Nothing is lost. The memory lives in Memocore, not in the tool, so a new client sees everything the old one knew the moment you connect it.
Yes. Create a company workspace, invite people by email and group memory into projects, inviting each person only into the ones they need — as editor or read-only viewer. See Memocore for teams for the details.
Yes. A team plan gives you API keys, and each key is limited to the projects you select for it. Your product searches that memory and can create memos in it too.
Yes, a free tier to start. Upgrade only when you outgrow it.
Still wondering about something? Ask the chatbot in the bottom-right corner — it knows Memocore inside out.
Give your AI a memory.
One memory, every AI client. Create your account — no password, just your email.
Free forever plan · set up in a minute