FAQ

Where does the inference come from?

Every call is served by the same frontier models you would reach going direct. Spoof runs a routing and aggregation layer in front of multiple model providers: rather than pay per-call list price, we source inference capacity in bulk and route each request down the most efficient path, then pass the wholesale rate through to you.

Is the response the real, unmodified model output?

Yes. What comes back is the model's genuine output, returned exactly as generated, byte for byte. Nothing is rewritten or injected.

Do I have an account or balance with Spoof?

You have your own wallet that only you control. Spoof never holds your funds, and the per-request charges you authorize can never exceed your balance.

Do I need to prepay or keep a balance with you?

No. You keep stablecoins in your own wallet and each call draws from it as you go. Nothing is deposited with Spoof.

Is each call really settled onchain?

Yes. Payments use x402, an open standard for settling onchain over HTTP, so every request is paid in a stablecoin straight from your wallet the moment you make it. There is no prepaid balance sitting with us.

Which chains and stablecoins can I fund with?

USDC on Base, Arbitrum, or Solana, or USDG on Robinhood. Billing follows whichever chain holds your balance. See Billing chains.

How is this cheaper than going direct?

Settling onchain removes the card processors and subscription margins that get baked into list prices, so you pay the metered rate with nothing added on top. We also run lean and source inference efficiently, then pass those savings straight through.

Do you store my prompts?

No. The API records nothing about your prompts or the model's responses. We meter usage for billing, never the content.

What if a request briefly fails?

Transient errors are retried and rerouted automatically, so a momentary hiccup almost never reaches you.