Your AI Agent Doesn't Need a $20/Month API Key.
It Needs 3 Cents and a Wallet.

Picture an AI agent doing regulatory monitoring. It's running on a schedule, and at 3 AM it hits a wall: it needs to check whether a country's crypto rules changed overnight. It has a task, it has a wallet, it has the ability to reason about whatever data it gets. What it doesn't have is an API key.

This is the structural gap nobody talks about. We've spent two years making agents smart and giving them tools. We haven't given them a payment rail that matches how they actually operate — bursty, autonomous, and metered in single requests rather than monthly seats.

A monthly subscription assumes a human signs up, a card gets charged, and a quota resets on the 1st. An agent doesn't work that way. It might fire once today and forty times tomorrow. It has no wallet to hand a card number, and you don't want it holding your card anyway.

Once an agent can pay per query, the workflows that were economically impossible become trivial. A few categories our users are building toward:

• Tourism & travel analytics — Pull inbound-arrival and seasonality figures for a set of Asian markets, then compare them in one reasoning pass. Instead of subscribing to five national statistics APIs, an agent asks five questions and pays for five answers.
• Regulatory & compliance monitoring — Track policy changes across Asia-Pacific and European jurisdictions on a schedule. The agent only pays on the runs where something is actually worth reporting, not a flat fee every month.
• Macro research — Ask for GDP, inflation, and unemployment figures across a basket of economies, then let the model compute the comparison. The data cost scales with the research, not with the calendar.
• Data-enrichment pipelines — Fill gaps in a dataset on demand. When a pipeline hits a missing country, it buys that one record instead of licensing the entire provider.

These are the kinds of workflows the platform is built to serve. Data availability and freshness vary by country and are visible through the free status endpoints before you pay for anything.

The core insight is simple: agents don't buy subscriptions, they buy answers.

Legacy APIs force a false choice — pay a monthly fee whether you make one query or ten thousand. But an agent doesn't want your whole API. It wants one slice of data to finish one task. The x402 protocol changes the unit of account from monthly subscription to per answer.

Three properties of the XRP Ledger make that economically viable rather than theoretical:

• Sub-cent precision — The smallest unit on XRPL is 1 drop (0.000001 XRP). You can price a single factual query and a deep AI analysis as different products without rounding either into meaninglessness.
• Fast settlement — XRP Ledger closes ledgers in a matter of seconds, so the payment is confirmed before the agent's next request. No billing job to catch up.
• The payment is the authentication — No API keys to issue, rotate, or leak. No rate-limit headers. No "upgrade your plan" wall. Pay, get data.

The free tier covers discovery — see which countries exist and how fresh their data is before spending anything. Paid queries settle on-chain.

XRP-denominated prices are fixed at the amounts above. Any USD equivalent depends on the live XRP market price at the time of payment and is not set by us.

# 1. Agent asks for data — anonymous, no key
curl -X POST https://api.xrplme.online/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
       "params":{"name":"xrplme_get_country_status","arguments":{}}}'

# → HTTP 402 Payment Required
# → response carries the price, a pay-to address, and an invoice ID

# 2. Agent pays that invoice (its wallet, its signature)
#    ... agent-side XRPL payment logic ...

# 3. Agent retries the SAME request with the payment proof
# → HTTP 200 — here's the data

The server verifies the payment on-chain and releases the data. There's no central billing database to drift out of sync and no token to silently expire mid-workflow — the ledger is the record.

The 402 response tells the agent exactly what to pay and to whom; nothing is hard-coded on the client. Working Python and JavaScript clients are in the MCP client repo.

1. Free, no wallet needed: query https://api.xrplme.online/mcp for the country list and status — no payment, no signup.
2. Add a wallet with a small XRP balance (Xaman, Joey, or any XRPL wallet).
3. Wire the loop: request → read the 402 → pay the invoice → retry. Grab a language example from the client repo.
4. Scale across all 30 countries — the same loop works for every market.

The dataset is 30 countries today, spanning Asia-Pacific, Europe, the Americas, and the Middle East. But the more interesting part isn't the dataset — it's the economic model.

Picture an agent that runs a research routine every morning: check a handful of markets, pay a few cents per data point, and only surface the findings that matter. Under a subscription model that routine costs the same whether it finds anything or not. Under micropayments it costs exactly what it used — which is what makes running it at all rational.

Need data from a country we don't cover yet, or want a data type exposed over MCP? Open an issue on GitHub — knowing which gaps matter is what decides where coverage grows next.