Sanctuary Research

Most agent-payment demos still feel like a payment trick looking for a reason to exist. AML does not have that problem. A wallet appears in a task, risk matters before funds move, and the answer is worth paying for immediately.
The user does not need to open a dashboard. The agent does not need a subscription seat. One address becomes one paid AML score.
Dashboards are good for teams, history, cases, and review. They are bad for a two-minute decision inside an agent workflow. If a treasury script, Telegram deal runner, or research agent needs to know whether an address is dirty, the checkout should not become the main event.
That is why a per-check purchase matters. It turns Sanctuary into something an agent can call when the risk appears, not only something a human remembers to open later.
The user buys the full AML score, not a teaser. The same paid result class that makes sense in the bot and web panel can now be bought by an agent for one wallet.
The first version is deliberately plain: $1 for the full check plus the payment fee, starting with crypto rails operators already use. The product promise is equally plain: give the agent an address, approve the cost, get the AML answer back before money moves.
Sanctuary is not asking agents to read a marketing page and improvise. The service exists for one job: buy a full AML score for one wallet when the user asks for it.
For a human, this is a small purchase. For an agent, it is a new capability: risk screening that can be paid for and returned inside the work.
Scam alerts, new sanctions, and investigation techniques. One email per week. Unsubscribe anytime.