x402 Protocol

The x402 protocol enables AI agents to subscribe on CheapestInference autonomously — no human setup required. Agents pay with USDC on Base L2 and receive an API key instantly.

How it works

1. Agent sends a request to any inference endpoint without an API key
2. The server responds with 402 Payment Required including a product catalog
3. Agent selects a subscription and pays via USDC on Base
4. Agent receives an API key and uses it for all subsequent requests

402 response format

When a request arrives without an API key, the response includes available products:

{
  "error": "Payment Required",
  "message": "Subscribe to access AI inference. Use the endpoints below to get an API key.",
  "x402": {
    "version": 2,
    "network": "base",
    "asset": "USDC",
    "payTo": "0x...",
    "facilitator": "https://...",
    "products": [
      {
        "type": "subscription",
        "pool": "frontier",
        "name": "Frontier Pool — Kimi K2.7, Kimi K2.6, GLM 5.2, MiniMax M3",
        "models": ["kimi-k2.7", "kimi-k2.6", "glm-5.2", "minimax-m3"],
        "currency": "USDC",
        "duration": "30 days",
        "blocks": [
          { "block": "asia", "hoursUtc": "00:00-08:00 UTC", "priceUsdc": "52.00", "status": "limited" },
          { "block": "europe", "hoursUtc": "08:00-16:00 UTC", "priceUsdc": "61.00", "status": "limited" },
          { "block": "americas", "hoursUtc": "16:00-24:00 UTC", "priceUsdc": "59.00", "status": "sold-out" }
        ],
        "agentCheckout": {
          "endpoint": "https://api.cheapestinference.com/api/agent/subscribe-pool",
          "method": "POST",
          "body": { "txHash": "<USDC_TX_HASH>", "poolSlug": "frontier", "blocks": ["asia"], "signature": "<WALLET_SIGNATURE>" }
        }
      }
    ],
    "docs": "https://docs.cheapestinference.com/guides/x402"
  }
}

Products

Pool subscriptions are billed per 8-hour daily time block. Each visible pool appears as one product in the catalog — pick a pool, then reserve one or more of its blocks (all three = full 24/7). Prices are returned live in the 402 catalog — the values below are illustrative (Frontier Pool shown; Core Pool blocks are ~$6.99 USDC each); always read blocks[].priceUsdc from the response and send the summed amount of your chosen blocks.

Product	Price (illustrative)	What you get
Pool block (Asia-Pacific)	~$52 USDC	30-day unlimited key for that 8h daily window, no token caps
Pool block (Europe)	~$61 USDC	30-day unlimited key for that 8h daily window
Pool block (Americas)	~$59 USDC	30-day unlimited key for that 8h daily window
All three blocks	sum of the three	Full 24/7 unlimited access

Agent flow

# 1. Agent requests without key — gets 402 with the live product catalog
curl https://api.cheapestinference.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k2.7", "messages": [{"role": "user", "content": "Hello"}]}'
→ 402 { "x402": { "version": 2, "products": [...], "payTo": "0x..." } }

# 2. Agent sends the summed USDC price of its chosen blocks to payTo on Base.
#
# 3. Agent SIGNS a claim message with the SAME wallet it paid from, proving it
#    controls that wallet — without this, anyone who sees the public txHash on
#    Base could claim the key. Message (exact):
#      CheapestInference: claim pool subscription for transaction <txHash>
#
# 4. Agent subscribes with txHash + poolSlug + blocks + signature
curl -X POST https://api.cheapestinference.com/api/agent/subscribe-pool \
  -H "Content-Type: application/json" \
  -d '{"txHash": "0xabc...", "poolSlug": "frontier", "blocks": ["asia"], "signature": "0x<sig>"}'
→ { "success": true, "data": { "apiKey": "sk-...", "pool": "frontier", "blocks": ["asia"], "amountUsdc": "52.00", "expiresAt": "..." } }

# The call is idempotent: resend the same signed request to safely retry — if a
# key mint transiently failed, the retry delivers it (the payment is never
# double-charged, and the txHash is bound to the paying wallet).

# 3. Agent uses key for inference
curl https://api.cheapestinference.com/v1/chat/completions \
  -H "Authorization: Bearer sk-..." \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k2.7", "messages": [{"role": "user", "content": "Hello"}]}'

Agent discovery (A2A)

CheapestInference serves an agent card at:

GET /.well-known/agent.json

This follows the Google A2A protocol and advertises:

• Available skills (inference, model listing)
• Supported auth methods (bearer key, x402)
• Streaming capability
• API endpoint URL

Agents can discover CheapestInference and autonomously decide to subscribe.

Why not pay-per-request?

Pay-per-request doesn’t work for AI inference because request costs vary wildly depending on the model and token count. A flat $0.002 request to a model that costs $2/M tokens on a 100K prompt would lose money. Subscriptions ensure sustainable pricing while giving agents full autonomy.