Frequently asked questions
The difference, the work, pricing and lifecycle, and how to get started.
The difference
How is this different from ChatGPT or Claude's memory?
Their memory is walled to their own app — you can't read inside it, version it, or take it with you. yarnnn is one memory that lives outside any single AI: every model reads and writes it, every change is signed and dated, and it's yours to export. Write it in Claude, and it's there in ChatGPT.
Is my data mine?
Yes. Every file is attributed, every version is kept, and the whole thing is exportable and reachable from any AI over MCP. It's yours to own — and we never train on it.
What is 'trace'?
Trace shows how any fact changed over time — who changed it, when, and what it was before. It's the thing a plain storage connector or an app's built-in memory can't show you: the full history behind what your AI knows.
Does it work across my team, not just me?
Yes — invite teammates by email and they join the same shared workspace. You, your people, and your AIs all write the same memory, every change signed with its author's name, and you can narrow or revoke anyone's access at any time. Working solo is just the simplest case.
The work
How do I put things in, and get them out?
Tell any connected AI to remember something, upload your files and notes, or connect the tools you already use (Slack, Notion). Any other AI — or teammate — can then recall it on the next session — no copy-paste, no re-explaining who you are.
Which AI models does it work with?
Any that speak MCP — ChatGPT, Claude, and others. It's neutral on purpose: it isn't tied to any one model, which is exactly why it can sit across all of them.
What's the 'second set of eyes' I've seen mentioned?
That's the optional checker (in beta): an assistant that reviews important work before it goes out, against rules you set, and keeps a record of every call it makes. The memory is valuable on its own; the checker is an upgrade you turn on when you're ready.
Pricing & lifecycle
What does it cost?
Your memory is free forever for one person — your files, your context, reachable from any AI. Pricing has two axes: a seat per teammate, and shared usage. The first seat (you, the owner) is free; each additional human is a paid seat ($20/mo), and AI connections are always free — never a seat. The paid plan includes a monthly shared usage pool the whole workspace draws from (owner-funded); heavier months you top up any amount from $5 (top-ups never expire). A solo workspace is free (usage-only); a team is paid per additional person. Every workspace starts with a $3 balance to feel the loop before you spend a cent.
Can I cap what it spends?
Yes — separately from your plan. Set a monthly budget ceiling and the assistant paces its own work to stay under it. That's a cap you set, not a charge we bill. Two guardrails, so spend is never a surprise: a budget ceiling you plan, and a floor (zero allowance and balance) that pauses the operation without losing anything.
What if my allowance runs out, or I turn the assistant off?
Nothing is deleted. When your monthly allowance and balance are spent, the assistant pauses — upgrade your plan or top up to resume. Turn it off and it simply stops drawing usage. Either way, your memory and every file stay yours, free.
Getting started
How do I start?
Start free. Connect an AI tool over MCP (or upload files, or just start typing), tell it something worth keeping, and watch it show up in the next AI you open — with the full history intact.
What's the best first move?
Connect the two AIs you use most, and save one thing in one of them. Open the other and recall it. That round-trip — write once, there everywhere, fully traceable — is the whole idea in about thirty seconds.
Still have questions?
Start free — save one thing in one AI and watch it show up in the next.