Nawame
FR · EN Get started
Working memory for your AIs

Your AI finally knows you — and won't forget.

One place where the context of your work lives. Your decisions, your notes, the way you do things. Every AI plugs in and picks up where you left off.

Start free See how it works →

no card · 2 minutes · Claude, Gemini, ChatGPT, Mistral

Today

You spend more time re-explaining than moving forward.

It's not a flaw in the models. They simply have nowhere to keep what you've taught them.

01

The conversation fills up

Three hours of work in one thread, then the window is full. You reopen somewhere else, and it's all to do again.

02

The morning copy-paste

The same block of context pasted back every day, a little different each time, never quite up to date.

03

Every AI starts from scratch

What you explained to one, the next one has no idea. Switching tools costs you half a day.

04

The lost decision

“We settled that back in March.” Finding where, when and why takes longer than deciding it all over again.

Nawame keeps all of it in one place, for every AI you use. See how it works →
How it works

Three steps, once and for all.

Nothing to install, nothing to tidy up. You keep talking to your AI exactly as usual.

01

You connect your AI

One line to paste into Claude, Gemini or ChatGPT. No install, no settings to figure out.

02

It remembers

As you go, it saves what matters: decisions, notes, ways of doing things. You write none of it yourself.

03

You carry on

Tomorrow, elsewhere, with another AI: it rereads the thread and picks the conversation up where you left it.

Every AI, the same memory

You switch AIs. Your memory doesn't.

Nawame connects through the open protocol these tools already share. What you teach one, the others know.

See how to set it up →
Clauderead and write
Geminiread and write
ChatGPTread and write
Mistralread and write
Your own toolopen API
For teams

What you know stops living in one head.

Share one folder and every AI on the team works from the same context. Onboarding someone new takes a conversation, not three weeks.

Share a folder

With one person or the whole organisation, to read or to write.

Seal what stands as truth

Binding decisions are marked: no AI can contradict them without citing them.

See who wrote what

Every version is dated and signed, human or AI. Nothing gets erased.

Sovereignty

Your memory belongs to you.

European hosting, encryption at rest, full export whenever you want. Your content is never used to train a model — ours or anyone else's.

Hosting

France · European Union

Encryption

at rest and in transit

Export

Markdown and JSON, any time

Training

never on your content

Access log

every AI read is traced

Pricing

Free to start, and it stays that way.

The Free tier is free with no time limit. Paid tiers will open later, for heavy use and teams.

Freeentry point
€0forever, no card required
  • Personal memory with no time limit
  • All your AIs connected at once
  • Nested folders and search
  • Markdown and JSON export
Start free
Powersoon
pricing soon
  • Everything in Free
  • Extended volume and advanced search
  • Long history, version comparison
  • Priority support
Soon
Teamsoon
pricing soon
  • Everything in Power
  • Organisations, roles and sharing
  • Sealed decisions, access log
  • Central billing
Soon

No payment is required today. Prices will be announced before anything opens.

Frequently asked

What people ask before they start.

Eleven short answers. If yours isn't here, we'd like to hear it.

Does Nawame replace my AI?
No. Nawame doesn't talk, answer or generate anything. It's the notebook your AI reads from. You keep using ChatGPT, Claude or Gemini exactly as before, now with a memory that lasts.
Which AIs does Nawame work with?
Claude, Gemini, ChatGPT and Mistral today, for both reading and writing. Nawame uses the open protocol these tools already share, and an open API lets you plug in your own tool.
Can I take my memory from one AI to another?
Yes — that's where the product starts. Your memory lives with you, not inside a model: you can switch AIs overnight without re-explaining a thing, and use several at once on the same context.
How is this different from the built-in memory in ChatGPT or Claude?
Built-in memory stays inside the tool that created it, it's opaque and you don't control it. Nawame's memory is visible, organised into folders, versioned, shareable, exportable, and readable by any AI.
Is my data safe? Who can see it?
Everything is private by default. You alone see your folders until you share something, and every share states with whom, to read or to write. Content is encrypted at rest and in transit.
Where is my data hosted?
In Europe, on French infrastructure. No copy is ever transferred outside the European Union.
Does Nawame access my private conversations?
No. Nawame only sees what your AI puts there, at your request. We don't read your content and it is never used to train a model, ours or anyone else's. Every read by an AI is written to a log you can review.
How does it actually work? What do I have to install?
Nothing to install. You create an account with your email address, paste one connection line into your AI, and that's it. Two minutes.
Is it really free?
Yes. The Free tier has no time limit and needs no card: personal memory, all your AIs connected, export included. Paid tiers will open later for heavy use and teams.
If I stop, can I get my data back or delete it?
You export everything to Markdown or JSON at any time, no need to ask. Deleting your account erases all of your content.
Does it work for a team too?
Yes. A shared folder gives every AI on the team the same context, with reader or actor roles, sealed decisions that stand as truth, and a signed history.