You can start from nothing. From then on every decision, constraint and useful detail stays in your Cobrain: Markdown notes that Claude, ChatGPT, Claude Code and any other compatible AI can read and update.
You switch AI, open a new chat, bring in a colleague: every time you have to rebuild the context. Decisions scatter across conversations and people, and the work starts again from the beginning.
It is not that the models have no memory: it is that the memory is locked inside somebody else's product. You cannot see what is in it, you cannot correct it, you cannot take it with you.
You connect your AI once. Whenever something comes up that will matter tomorrow too — a decision, a constraint, a contact — it can save it in your Cobrain. In the next conversation it is already there, even if you switch AI.
And everything stays open. They are Markdown files in real folders: you open one, fix a line, move a note, delete what you no longer want the AI to know.
Share a folder, or even a single note, with a colleague. It shows up inside their brain and becomes context for their AI too.
The client, their decisions, their constraints: you work on them with Claude, your partner with ChatGPT, both of you on the same notes. Neither of you starts over.
Folders on the left, the note in the middle, on the right who wrote it and when. The same things your AI sees.
Keep using whichever AI you prefer. You ask the same question as always; it finds in your Cobrain the context you used to have to explain again.
The two questions people ask before signing up: what it costs the AI to read the brain, and who else can read it.
15×fewer tokens to start a chat
Without memory, every new chat means explaining it all again or pasting the project documents, and the AI rereads them in full. With Cobrain it starts with the rules, the card and the latest decisions; for the rest it searches, and opens only the notes it needs.
On the APIs and pay-as-you-go plans that is money. On a subscription that counts usage, like Claude's, it means reaching the limit later.
AES-256one key per brain
Notes, versions and attachments are encrypted with their brain's key. The database and the file store hold only encrypted text: the database cannot read your notes, and whoever stores your files cannot open them.
The keys are held by AWS KMS and only Cobrain's server uses them, to show the notes to you and your AI. Every time a key is opened, it is logged. Note and folder names, tags and dates stay readable: they are what the tree is built from.
Your AI opens your files and works with your permissions, the same as the site's: if you cannot see or change something, neither can it.
PDFs, Word, Excel, photos: attach them to a note and your AI reads them, Claude or ChatGPT in a chat included. Nothing to download and re-upload in every conversation. On the Pro plan.
Not by anyone, and not by any AI. You choose who comes in, and whether they can only read or write too.
Each one works with your permissions: it sees nothing you would not see.
Who changed what, and when — including when it was your AI. A wrong edit can be undone.
Headings, lists, quotes, code. Underneath they stay .md files: no format only we can read.
An archive of folders and files. It opens with Obsidian, with a text editor, with anything. The data sits in Europe, in Ireland.
The free plan is not a trial: you can genuinely build your brain on it. You move to Pro when you want to share and work on the same notes with other people.
We do not ask for a card, and it does not expire.
Only the sharer pays, never the receiver.
Made for agencies, studios and small teams. Clients and collaborators who receive a branch still come in free.
Line by line, Team included: all the plans compared.
What people usually ask us before signing up.
Claude, ChatGPT, Claude Code, Codex and any other AI that supports MCP. You paste an address and authorise once. Today Claude requires a paid plan for custom connectors; on ChatGPT you upload our plugin, a file you download, on any plan. Those are their requirements, not ours.
No. Most people start from nothing and their Cobrain grows as they work. If you already have notes somewhere — in Notion, in a folder, in old chats or in Obsidian — your AI can bring them in: there is a guide.
Yes, on the Pro plan. Attach a file to a note and your AI opens it: PDFs, scanned ones included, Word, Excel, PowerPoint, Pages and Keynote. From a video it sees a few frames; audio, not yet.
Never. A free account is enough. The plan is paid for by whoever shares.
You lose nothing. All your notes stay, and shares already active keep working read-only. You go back to the free plan limits for anything new. Reactivate and everything picks up where it was.
Yes, whenever you like, in one go. They are .md files in folders: what comes out opens with anything.
On servers in Europe, in Ireland. Notes and attachments are encrypted with one key per brain: the database cannot read them. Only Cobrain’s server opens them, to show them to you and your AI. Every save keeps the previous version.
It is free and no card is needed. You sign in with a link sent by email, connect your AI and start building a memory that does not stay locked inside a chat.