Use Agent402 as your OpenClaw model provider - pay by card, no wallet
OpenClaw talks to any OpenAI-compatible provider through
one block in openclaw.json. Agent402's LLM gateway is one of those, with a
twist: it can be paid by card, through a prepaid credits key, so an agent
runs without a crypto wallet. USDC over x402 works too, if you'd rather.
1. Get a credits key (card, two minutes)
Buy a pack at agent402.tools/credits by card.
The key (a402_…) is shown once on the thanks page and emailed. It spends on
any paid route, including every gateway tier below, and
GET /api/credits/balance (Bearer) reports what is left.
Put it in the environment OpenClaw runs in:
export AGENT402_CREDITS_KEY=a402_…
2. Add the provider
// ~/.openclaw/openclaw.json
{
agents: {
defaults: {
model: { primary: "agent402/auto" },
},
},
models: {
providers: {
agent402: {
baseUrl: "https://agent402.tools/v1/auto",
apiKey: "${AGENT402_CREDITS_KEY}",
api: "openai-completions",
timeoutSeconds: 120,
models: [
{
id: "auto",
name: "Agent402 auto (routed, $0.01/call)",
reasoning: false,
input: ["text"],
// Agent402 bills a flat $0.01 per call, not per token, so
// OpenClaw's per-token cost display does not apply and stays at zero.
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
contextWindow: 4000,
maxTokens: 1024,
},
],
},
},
},
}
Restart the gateway (openclaw gateway restart). Every model call now goes to
POST /v1/auto/chat/completions with your credits key, and auto picks the
model per prompt (code, reasoning, long-context or general) from a ranked
table, with failover if the first choice is down.
Explicit models and the other tiers
The gateway has 6 chat tiers, each a flat price per call:
| tier | baseUrl path | per call | max output tokens |
|---|---|---|---|
| nano | /v1/nano |
$0.003 | 768 |
| auto (routed) | /v1/auto |
$0.01 | 1024 |
| base | /v1 |
$0.02 | 2048 |
| grounded (web search on every call) | /v1/grounded |
$0.03 | 1024 |
| pro | /v1/pro |
$0.10 | 4096 |
| premium | /v1/premium |
$0.50 | 8192 |
To pin a model, add a second provider whose baseUrl is that tier's path and
whose models[] list ids from /v1/models,
for example baseUrl: "https://agent402.tools/v1/premium" with
{ id: "openai/gpt-5" } and { id: "anthropic/claude-opus-5" }. A model sent
to the wrong tier is answered with a 400 that names its home tier; nothing is
charged.
Pay from a wallet instead
Every tier answers an x402 402 (USDC on Base, Solana, Polygon, Arbitrum, Monad, Celo, Avalanche, Sei, Optimism, Stellar, or Algorand - or USDG on Robinhood Chain) and an MPP challenge, so any
x402-capable client can pay per call with no key at all. For OpenClaw that
means a local proxy that signs the payment and forwards the request, the same
pattern other x402 routers use; agent402-client already does the paying half
(x402, MPP, credits, free tier), and an OpenClaw plugin that wraps it as a
provider is in progress. Until it ships, the credits-key block above is the
supported path.
What you get that a plain router does not
The same key and the same base URL reach the rest of the catalog: 500+ deterministic tools (search, extract, render, PDF, EDGAR, openFDA, on-chain data), the smart order router that pays other x402 sellers on your agent's behalf, and receipts for every call. Prices on this page are read from the live gateway configuration.