An AI in your account should be able to read everything and change nothing.
That’s the whole idea. Connect the assistant you already pay for, let it work inside your real records, and keep every action that matters behind your own approval.
What a connected AI client can read
It only sees what falls inside the scopes you grant. Nothing account-wide happens by default.
What it can never do on its own
Reading a record and acting on it are two different permissions. Sending a message to a seller or contractor, signing or generating a document, moving money, locking a comp set or a scope of work — every one of those is a write, and every write stages as a draft first.
How permissions actually work
Under the hood, rei.OS uses scoped connections instead of all-or-nothing access — a client can hold something like comps:read without holding scope:write, so it can pull comps for you but can't touch your scope of work.
Your bank connection, specifically
rei.OS connects to your bank through Plaid — the same connection layer used by a large share of banking and finance apps you probably already use. That means rei.OS never sees or stores your bank username or password. Plaid authenticates you directly with your bank and passes rei.OS only the account and transaction data you approve.
Where your data lives
Your records sit on securely hosted infrastructure, encrypted both in transit and at rest. Formal third-party security certification is on our roadmap, not a claim we're making today — we'd rather tell you exactly where things stand than round up.
If you ever want out
Every connected client — Claude, ChatGPT, anything added later — can be revoked from Settings > Connected Clients at any time. Revocation is immediate: access ends the moment you click it, not at the end of a billing cycle.
— Bryson, Founder, rei.OS
REI.OS LLC
Nothing sends, signs, or pays without your approval.
See it on your own records. Thirty days free, no card.