Skip to main content
Your agent has a wallet. Now it needs to use it. This page covers two key concepts: how your agent pays for services automatically using the x402 protocol, and how skills define the actions your agent can perform.

​
x402: pay-per-request with stablecoins

x402 lets your agent pay for API requests using without subscriptions or API keys. Your agent requests a resource, the server tells it the price, your agent pays, and the server delivers the data. The name comes from HTTP status code 402 Payment Required, which has been reserved in the HTTP specification since the 1990s but was never widely adopted. x402 finally puts it to use.

​
How x402 works

1

Your agent makes a request

Your agent sends a standard HTTP request to an API endpoint, just like any other API call.
2

The server responds with 402 Payment Required

Instead of returning data, the server responds with a 402 status code and includes payment details: how much it costs, which stablecoin to pay in, and where to send the payment.
3

Your agent pays

Your agent’s wallet constructs and sends a onchain payment based on the server’s instructions. This happens automatically, no human approval needed.
4

The server delivers the data

Once the payment is confirmed, the server returns the requested data. The entire flow takes seconds.
This model is powerful for agents because it requires no pre-existing relationship between your agent and the service. Any agent with a funded wallet can pay for any x402-enabled API on the fly. Learn more about x402 →

​
Skills: actions your agent can perform

Skills are typed function definitions that describe an action your agent can invoke. Each skill specifies a name, expected inputs, and return type similar to a tool or function schema in any LLM tool-use integration. Your agent parses these definitions at runtime and calls them when relevant. Skills cover actions like checking a token balance, executing a swap, sending a payment, or calling a .

​
What a skill looks like

A skill definition includes the action name, a description of what it does, the required inputs, and the expected output. Here’s a simplified example:
Title
{
  "skill": "check-balance",
  "description": "Check the token balance of a wallet address",
  "inputs": {
    "wallet_address": "The address to check",
    "token": "The token symbol (e.g., USDC, ETH)"
  },
  "output": "The current balance of the specified token"
}
Your agent reads this definition, understands what inputs it needs to provide, and calls the skill when it’s relevant to the task at hand.

​
Skill providers

What they are: Pre-built skills for agents running on OpenClaw. BANKR skills cover common actions like token swaps, balance checks, and DeFi interactions.Best for: Agents using OpenClaw or BANKR that need ready-to-use onchain capabilities.Browse BANKR skills →

​
Next step

Identity, verification, and auth

Make your agent discoverable and verifiable so other agents and services can trust it.