Tamper-proof audit trails for AI agents. Verifiable by anyone. Alterable by no one.
Writes cryptographic fingerprints of AI agent actions to Base L2 blockchain. Only the hash goes on-chain — no PII, no sensitive data. Includes TypeScript and Python SDKs, a CLI verifier, and a Next.js dashboard.
- On-chain audit trail on Base L2 (Ethereum)
- Model version pinning (EU AI Act compliance)
- Dead man's switch for contingency triggers
- TypeScript SDK (31 tests)
- Python SDK (31 tests)
- CLI verifier tool
- Next.js dashboard (live on Vercel)
npm install
npx hardhat compilecd sdk && npm installcd python && pip install -e .import { ChainLog } from '@chainlog/sdk';
const cl = new ChainLog({ rpcUrl, contractAddress });
await cl.log({ action: 'inference', model: 'claude-sonnet-4-6', hash: '...' });from chainlog import ChainLogClient
cl = ChainLogClient(rpc_url=..., contract_address=...)
cl.log(action="inference", model="claude-sonnet-4-6", hash="...")npx hardhat test # Contract tests
cd sdk && npm test # TS SDK tests
cd python && pytest # Python SDK tests
cd cli && npm test # CLI testsSprint 6 complete. Dashboard live on Vercel. Contracts pending mainnet deploy (needs funded Base Sepolia wallet).
Proprietary