x402 validation layer

Make every API response safe for agents.

AgentProof validates, repairs, and scores structured API responses before an AI agent acts on them.

Validate response Read docs

98/100 Example Proof Score for a repaired response.

4 Simple paid endpoints ready for x402 agents.

$0.001+ Micropriced checks that can run before an agent acts.

JSON

Schema

Repair

Score

Agent

Trust report before execution.

Agents get a compact proof object: schema match, issues found, suggested repair, and a score they can use before taking the next step.

Repair suggested Schema mismatch caught before the agent acts.

$0.002

{
  "service": "AgentProof",
  "valid": false,
  "errors": [{ "path": "/user_id", "expected": "string" }],
  "repairSuggestion": { "user_id": "123" }
}

Small checks agents can buy.

Each endpoint is narrow, fast, and useful as a paid guardrail in an automated workflow.

POST/json/repair$0.001

POST/schema/infer$0.002

POST/api/validate$0.002

POST/api/contract-score$0.003

AgentCash ready

npx -y agentcash@latest fetch https://agentproof-murex.vercel.app/api/validate \
  --payment-protocol x402 \
  --payment-network base \
  --max-amount 0.002

Built for agent workflows, not dashboards.

AgentProof focuses on the tiny moment between receiving external data and letting an AI agent act on it.

Validate Catch schema drift early.

Compare an API response against the contract the agent expected and return precise, machine-readable errors.

Repair Return a safer next payload.

Suggest a corrected JSON object when the response can be repaired without external data.

Score Give agents a simple confidence signal.

Turn repeated samples into a contract score that helps agents decide whether to continue, retry, or escalate.

Test AgentProof on a real local response.

Use the browser demo for product review, or open the docs for x402, curl, and AgentCash integration examples.

Open live demo Open docs