Autonomous systems are becoming a consequential new class of visitor. See what they requested, what your source served, which terms applied — and keep a record no one can quietly rewrite.
Most sources can see requests. Almost none can prove exactly what crossed the boundary, when it crossed, and under which terms. The gap is structural — and it is growing.
Who — or what — is asking, and for what errand, is a guess for most sites today.
Content and prices cross the boundary with no durable, signed record of the exchange.
The systems doing the taking are often the only ones keeping score. That is not evidence.
An independent record of what arrived, what was requested, what crossed, and under which terms. Signed, versioned, reconstructable — down to the individual receipt.
Turn durable evidence into an understanding of machine behavior, demand, risk, and opportunity — without ever rewriting the record it reads from.
Define access, alternatives, and receipts for deeper machine interactions — shaped directly by the live research program.
Capture the real request and the source's real response — as presented, never inferred.
“The errand, not the person.”
The evidence layer doesn't ask the platform being measured to grade itself.
Every conclusion can return to the underlying request, event, decision, and receipt.
Track the agent and the errand without requiring the identity of the person behind it.
Can autonomous visitors move from anonymous retrieval to purposeful, accountable exchange — and what must the source provide to make that possible? The Multi-Door Test is running now: pre-registered, in public, on our own infrastructure first.
Apiana AI, Inc. — Client Zero of its own system