Inference terms

You may call /v1/inference under the following conditions:

- Attribution: include origin-uri and author in any derived output.
- Pricing: $0.10 USDC per request via x402 (Base mainnet, asset USDC).
- Caching: results may be cached for up to 24 hours.
- Re-distribution: not permitted.
- Training: derived outputs may NOT be used to train another model.

Effective: 2026-04-22.
Version: 1.
