# ProofGate x402

Deterministic safety scanning for prompts and agent tasks before execution.

## Endpoints

- Free preview: POST https://proofgate-agent-safety.arinagame1993.chatgpt.site/v1/preview (JSON {"text":"..."}, up to 512 UTF-8 bytes).
- Paid scan: POST https://proofgate-agent-safety.arinagame1993.chatgpt.site/v1/scan (JSON {"text":"..."}, up to 16 KiB UTF-8 bytes).
- OpenAPI: https://proofgate-agent-safety.arinagame1993.chatgpt.site/openapi.json
- Agent manifest: https://proofgate-agent-safety.arinagame1993.chatgpt.site/.well-known/agent.json

## Payment

- Protocol: x402 v2, exact scheme
- Price: $0.001 USDC per successful paid request
- Network: Base mainnet (eip155:8453)
- USDC asset: 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913
- Pay to: 0x56262c850df2cbdc11e8f48a38385ebf316691ca
- Facilitator: https://facilitator.xpay.sh

Send the request once without a payment header to receive the standard HTTP 402 payment requirements. Retry with a valid PAYMENT-SIGNATURE header. The result is released only after verify and settle succeed.

## Output

The paid response contains score, severity, categories, rule-based findings with character spans, and safe actions. It never includes or stores the submitted text.

## Safety and privacy

No LLM calls, content URL fetching, shell execution, credential collection, request-body retention, or request-content logging.