Pink Agentic AI Payment lets AI agents connect and pay through MCP while a policy engine enforces each customer's spending rules before any payment executes.
Join early accessThree steps. After the one-time MCP connection, policies are designed to be managed in the console, not in code.
Illustrative examples of the kind of rules a customer could configure in the console. These are examples of the product's intent, not screenshots of a shipped UI.
| Agent | Spending limit | Human approval required | Scope |
|---|---|---|---|
| procurement-bot | Up to $500/day | Above $200 | SaaS vendors on the allowlist only |
| research-assistant | Up to $50 per transaction | — | Data and API providers; no spend on advertising or crypto categories, ever |
| travel-booker | Up to $2,000 per trip | Anything outside the travel category (auto-denied) | Flights and hotels for pre-approved employees |
Illustrative — API not yet public. Field names, transport, and syntax are subject to change.
{
"mcpServers": {
"pink-agentic-payment": {
"url": "<PINK_MCP_SERVER_URL>",
"headers": {
"Authorization": "Bearer {{AGENT_API_KEY}}"
}
}
}
} Once connected, an agent would call a payment tool exposed by the server (for example, check_payment or request_payment); the server evaluates the request against the console's active policy before allowing it to proceed. Exact tool names and request/response shapes are not finalized.
No. It's in early access. Join the waitlist and tell us your use case to be considered.
MCP (Model Context Protocol) is an open standard for connecting AI applications to external tools and data sources. We use it so any MCP-compatible agent can connect to our payment and policy layer the same way it connects to other tools, instead of requiring a custom integration.
No — those protocols define how a payment is authorized and executed at the network or checkout level. We sit at the policy layer: deciding whether a given agent is allowed to make a given payment, and enforcing that before execution.
Not yet decided publicly. We're evaluating options and will share details as the product matures — early-access feedback will help shape it.
That's the core design goal we're building toward: policies are enforced by the console/server, independent of what the agent itself was instructed to do, so a misconfigured or misbehaving agent request is still checked against your rules. We'll publish specifics on enforcement guarantees as the product nears release.
The customer. The console is built for your finance, procurement, or IT team to define and change policies without engineering involvement.
Pink Agentic AI Payment is a new, separate product; PinkWallet's existing payments and compliance experience informs how we're building it, but the two are not the same offering.
Join the early access waitlist. We're prioritizing developers and companies with a concrete agent-payments use case they can describe to us.
Definitions of the protocols and terms above, and our original research on where agentic payments actually stand today.
Contact Email: [email protected]
ALTIVOX SPHERE SPÓŁKA ZO.O. is a wholly owned subsidiary of PinkWallet, registered in Poland in the Polish Ministry of Finance's virtual-currency activity register (registration number: RDWW-1797), authorized to provide cryptocurrency exchange and digital asset custody services in the EU under AML/CTF regulations.
PinkWallet (Pink Technologies Inc) is registered as a Money Services Business with the U.S. Department of the Treasury, FinCEN (Registration No. 31000321450011), enabling it to provide users with cross-border fund settlement and international payment services under the U.S. regulatory framework.
PinkWallet (Pink Wallet Technology Inc.) is registered as a Money Services Business with FINTRAC, Canada's financial intelligence unit (Registration No. C100001073), enabling it to provide users with cross-border fund settlement and virtual currency services under the Canadian regulatory framework.
