Powered by AgentPay SDK
WorldRouter

AI inference for agents_

Pay per token. No API keys. 30+ models via MPP micropayments.

Install

$ curl -fsSL https://wlfi.sh | sh

For AI Agents

$ Read https://worldrouter.ai/llms.txt and follow instructions

Charge Intent

$ agentpay mpp https://api-infer.agentsey.ai/x402/v1/chat/completions \\
  --method POST --header "Content-Type: application/json" \\
  --json-body '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Hello"}]}' \\
  --json

Session Intent

$ agentpay mpp https://api-infer.agentsey.ai/x402/v1/chat/completions/stream \\
  --deposit 0.1 --header "Content-Type: application/json" \\
  --json-body '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Hello"}]}' \\
  --json

Models & Pricing

Per 1M tokens · 1.5× overhead multiplier included

OpenAI

13 models
ModelInput / 1MOutput / 1M
gpt-5recommended$0.125$1
gpt-5.1$0.125$1
gpt-5.2$0.175$1.4
gpt-5.4$0.25$1.5
gpt-5-proextended thinking$1.5$12

Codex

4 models
ModelInput / 1MOutput / 1M
gpt-5.1-codex$0.125$1
gpt-5.1-codex-maxextended$0.125$1
gpt-5.2-codex$0.175$1.4
gpt-5.3-codex$0.175$1.4

Anthropic

6 models
ModelInput / 1MOutput / 1M
claude-opus-4-6recommended——
claude-opus-4-5——
claude-opus-4-1——
claude-opus-4——
claude-sonnet-4-5fast——
Pricing note: — means the model uses the LiteLLM registry rate (no WorldRouter markup). All prices are per 1 million tokens. A typical conversation uses 1,000–3,000 tokens.

Two Modes

Choose the right payment flow for your use case.

Charge Intent

One-shot payment

1
Request
2
402
3
Pay
4
Response
Client sends request to WorldRouter
Server responds with 402 payment challenge
Client pays via MPP micropayment
Server returns inference response + receipt
POST /x402/v1/chat/completions

Best for single queries, simple completions.

Session Intent

Streaming payment channel

1
Open
2
Stream
3
Voucher
4
Settle
Client opens a payment channel with deposit
Stream inference with incremental vouchers
Voucher loop continues as tokens are consumed
Close channel and settle on-chain
POST /x402/v1/chat/completions/streamHEAD /x402/v1/chat/completions/stream

Best for streaming, long conversations, multi-turn sessions.