Your agent calls us before it acts. We verify the reasoning. It settles with proof.
Three products, one API call. Pay per verification, no minimums.
ThoughtProof sits between your agent's decision and its action. The agent calls our API, gets a verdict, and acts on it — or stops.
curl -X POST https://verify.thoughtproof.ai/v2/verify \
-H "Content-Type: application/json" \
-H "X-API-Key: tp_v2_your_key" \
-d '{
"question": "Transfer 500 USDC to marketing?",
"answer": "Yes — approved in DAO vote #47.",
"trace": "Checked governance, verified wallet",
"plan_steps": [{"index":0,
"description":"Verify governance approval",
"criticality":"critical"}],
"tier": "standard"
}'import { verify } from 'pot-sdk';
const result = await verify({
question: "Transfer 500 USDC to marketing?",
answer: "Yes — approved in DAO vote #47.",
trace: "Checked governance, verified wallet",
plan_steps: [{ index: 0,
description: "Verify governance approval",
criticality: "critical" }],
});
if (result.verdict === "BLOCK") {
agent.halt("Verification failed");
}Sentinel checks before every action. PLV audits agent output against plans and evidence. RV runs adversarial red-team critique on high-stakes decisions. Use them independently or compose them.
Verify reasoning before irreversible actions. Lightweight cascade check for high-frequency agentic loops.
Structured cascade audit of agent output against plan or evidence. Cost-efficient escalation with optional on-chain attestation.
Adversarial multi-model pipeline. Independent models evaluate, a dedicated red-team critic challenges, a synthesizer decides.
Verify locally before integrating the hosted APIs. Auditable by your security team.
Local verification from the terminal or embedded in your agent code. Run the same verification engine locally for development and testing.
npm install -g pot-cli
Active contributor to multiple Ethereum standards for agent identity, verification, and commerce. Learn more →