Agent API (x402)Pricing

Pricing

Rule: every paid route charges the Venice API cost equivalent × margin multiplier (X402_VENICE_PRICE_MULT).

Published price list

GET https://x402.aispace.bot/openapi.json → info.x-pricing:

  • Live Venice catalog rates × current priceMult
  • Per-route accuracy tags (estimate, list_rate, quote_required, …)
  • howToGetAccurateEstimates for chat / image / music / video / probe

Changing X402_VENICE_PRICE_MULT updates this list on the next OpenAPI fetch and all new 402 challenges.

Settlement rule: always pay live accepts[].amount. The list is for discovery and budgeting.

Accuracy

TagMeaningGet an accurate number
estimateDepends on tokens/chars/secondsSet max_tokens / duration / known size, or unpaid probe → read 402
list_rateFixed Venice unit × multUse model (+ resolution) from x-pricing.models
quote_required / reference_onlyNo stable job list priceFree POST /api/v1/video/quote or /audio/quote → { quote }
exact_via_server_quoteQueue billingServer re-quotes Venice; 402 = quote × mult
freeNo charge on this gateway—

Formula

agent charge = veniceUpstreamUsd × X402_VENICE_PRICE_MULT

Then apply min charge. If the result would exceed X402_VENICE_MAX_CHARGE_USD, the request fails closed (charge_exceeds_max).

Budget tips

  • Chat: set max_tokens when you can.
  • Music / video: free /quote first; /queue uses a fresh server-side Venice quote.
  • Start with enough USDC for real job sizes (video can be dollars).

Operator counters: GET https://x402.aispace.bot/api/x402/stats

Async music & videoErrors