Claude + GPT-4o in a multi-round consensus loop. Higher confidence than any single model. Pay per scan via x402 micropayments.
Two leading LLMs adversarially review your code, then reconcile their findings into a single high-confidence report.
POST your smart contract source to /scan with an x402 micropayment.
The first LLM performs deep vulnerability analysis, identifying potential issues with severity ratings.
The second LLM adversarially reviews the first LLM's findings, confirming, disputing, or adding new issues.
The first LLM reconciles disagreements. If models converge is high, results finalize. Otherwise, loop continues.
The API runs on port 8403. All endpoints below are relative
to the API base URL.
Deep vulnerability detection across smart contract and general-purpose languages.
Reentrancy, access control, flash loans, oracle manipulation, MEV, delegatecall injection, storage collision, signature replay
Missing signer checks, PDA seed collisions, CPI vulnerabilities
Resource safety, missing acquires annotations
Reentrancy, overflow, access control, storage layout issues
Injection, path traversal, deserialization, hardcoded secrets
XSS, prototype pollution, SSRF, ReDoS
All pricing is in USDT on TRON.