Snaplii
Create Account Need Help?
Agent-to-Merchants (A2M)

North America’s First
Real-World AI Payment

Others announced it, we shipped it. A working agent-to-merchant payment layer built on gift cards and the Prepaid Payment Network: MCP-native, pre-funded, and live across 500K+ locations today.

Callable from Claude OpenClaw MCP clients
client.py MCP-native
payment_ctx = {
    "specifiedPrimaryPaymentMethod": payment_method,
    "voucherOption": "BEST_FIT",
    "cashbackOption": "USE",
}
if payment_token:
    payment_ctx["specifiedPrimaryPaymentToken"] = payment_token
return self._post("/v2/purchase", json={
    "orderInfo": {
        "orderType": "GIFT_CARD",
        "item": {"itemId": item_id, "price": price},
        "orderContext": {"giftOrder": "false"},
        "businessChannel": "APP",
    },
    "paymentContext": payment_ctx,
    "delivery": {"type": "WALLET", "immediateSend": "true"},
    "locationProv": location_prov,
})
How it works

A controlled layer between user, agent, and merchant

Users fund Snaplii and set the boundary. Agents operate within it. Snaplii turns payment into merchant-ready value, no credentials exposed.

User
Funds Snaplii and sets the agent’s limits, methods, and scope.
Agent
Decides, navigates, and executes within its boundary.
Snaplii
Issues a one-time, transaction-scoped instrument. Real cards never exposed.
Merchant
Accepts it as a normal gift card. No integration, 100% success.

Isolated by design

Tokenized

Real card credentials are never exposed to the agent.

Scoped

Agents operate only within pre-approved limits and permissions.

Pre-funded

A hard cap that stops when the balance depletes, no overspend.

Merchant-bounded

Each credential is single-use and bound to the transaction.

500k+
North America locations, live today
100%
Payment success rate
0
Merchant integrations needed
Up to 15%
Cashback on every txn
Watch the demo

This isn’t a concept. It’s working.

An AI agent prepares payment via Snaplii, receives merchant-ready value, and completes a real DoorDash order within a controlled, pre-funded boundary.

a2m-muse-snaplii-demo.mp4
We don’t build AI agents. We make them pay.Watch more demo on Youtube
Two paths, side by side

Built on foundations merchants already accept

Card-network protocols are the long-term future. Snaplii completes real transactions on existing rails live today.

Other AI Payment
Snaplii A2M
Payment method
Tokenized card
Gift card + Prepaid Network
Per-transaction user confirmation
Required for every transaction
Not required — pre-funded & autonomous
Merchant protocol integration
Merchants must integrate the stack
None — merchants already accept gift cards
Large retailer coverage
Takes time to onboard
Amazon, Walmart, Home Depot supported
Offline / in-store support
Limited
Gas, restaurants, retail, coffee
Overspend risk
Token can be reused
None — pre-funded hard cap
Cashback rewards
None
Up to 15% on every transaction
Available today
Partially
Fully — open-source MCP server, live

Built for agent frameworks

Authorize the boundary, issue an API key, and let the agent execute. Compliance stays with humans; execution moves to agents.

MCP-native

Callable directly from Claude, OpenClaw, and other leading AI clients.

API key

Authorize spending limits, methods, and scope, then issue a key. Multiple keys supported per user.

Decide · navigate · execute

Agents do the work; Snaplii completes safe payment execution.

Start building agent payments

Open-source, MCP-native, and live across 500K+ locations today.

GitHub repo
Build your first agent payment in minutes.
Clawhub skill
snaplii-a2m-payment
MCP plugin
snaplii-a2m-mcp