Pickerbird enforces its guarantees in code rather than in a policy document. The limits you set are built into how the app works, which means they cannot be quietly rewritten later — not by an update, not by a change of ownership, not by us.
Almost every privacy policy is a promise about the future. We won’t look. We won’t sell it. We’ll keep it safe. The words are usually sincere. But a promise has a property worth noticing: the person who made it can change it. Policies get updated. Companies get bought. “We won’t” becomes “we now do,” in an email you skim and forget.
We didn’t want Pickerbird’s core promise to be that kind of promise. So we built it into the parts of the product that can’t be quietly rewritten later.
Here’s what that means in practice.
Can Pickerbird read my content?
No — and that’s a fact about the design, not a pledge.
Your messages and notes are sealed with keys we don’t hold. When something is waiting to be delivered, what passes through our systems is scrambled and unreadable to us; once it’s delivered, it’s gone. We’re not promising not to peek. We’ve built it so there’s nothing to peek at. The difference matters most on the day someone with authority — a new owner, a court, an ambitious executive — asks us to hand your content over. You can’t hand over what you can’t read.
Can anyone override my settings?
Nobody can — including us.
When you mark something Closed, that’s enforced by how the app is built. There’s no admin switch on our side that flips it open, no “for support purposes” back door. If we could override your rules, they wouldn’t be your rules; they’d be ours, lent to you until further notice. The setting itself is one of three access tiers you assign per item.
Can I see what my AI actually does?
Yes, and you shouldn’t have to take our word for it.
Trust shouldn’t require faith. Pickerbird keeps a plain record of what the assistant reached for, what it sent to whichever AI brain you chose, and what it held back — and that record is yours to open and to keep. Most products ask you to trust that the assistant behaved. We’d rather just show you.
What does this actually guarantee?
That we are less able to break the promise — which is the part you can rely on.
None of this makes us more virtuous than the next company. A guarantee isn’t a stronger promise. It’s a promise nobody — not even the person who made it — can take back.
That’s the whole idea behind Pickerbird: not an AI that swears it will behave, but one that’s built so it can’t do otherwise.
I worked through this argument in more detail here: Lock and Key against AI Agents.
Join the waitlist to be among the first to try Pickerbird.
FAQ
What does "enforced in code, not policy" mean?
A policy is a promise about the future, and the person who made it can change it later. Pickerbird instead builds its guarantees into the parts of the product that cannot be quietly rewritten, so the limits you set hold whether or not anyone is watching.
Can Pickerbird read my content?
No. Your messages and notes are sealed with keys we don't hold. What passes through our systems while something waits to be delivered is scrambled and unreadable to us, and once it's delivered it's gone. That's a fact about the design, not a pledge.
Can anyone override the settings I choose?
No — including us. When you mark something Closed, that's enforced by how the app is built. There is no admin switch on our side that flips it open and no "for support purposes" back door.
Can I see what the AI actually did?
Yes. Pickerbird keeps a plain record of what the assistant reached for, what it sent to whichever AI brain you chose, and what it held back — and that record is yours to open and to keep.