About Witz Labs

Our Mission

WITZ Guardian is the first platform specialized in governance, auditing, and containment of autonomous AI agents operating in Web3 infrastructure. We combine proprietary centralized intelligence with a distributed observation network to become the standard trust layer for any autonomous agent interacting with digital assets.

In 2026, the convergence of autonomous AI agents and decentralized finance has created a critical security gap. Models reach 72.2% success rates in exploitation scenarios. The ratio of agents to humans is 82:1 in organizations. Only 6% have an advanced AI security strategy. WITZ Guardian fills this gap.

What We Are

Think of us as the credit rating agency for autonomous AI agents. Just as Moody's and S&P evaluate companies, WITZ Guardian continuously evaluates agents with three complementary metrics:

• Risk Score — How risky is this agent right now? (dynamic)
• Reputation Score — How trustworthy has this agent been over time? (cumulative)
• KYA Certificate — Who is this agent and is it verified? (on-chain NFT)

The difference: our scoring is continuous (not point-in-time), on-chain (verifiable by anyone), and powered by proprietary AI (not human opinion).

What We Are NOT

• Not a smart contract auditor (we audit agent behavior, not code)
• Not a firewall (we monitor and alert, not block network traffic)
• Not a custodian (we never control or have access to funds)
• Not a replacement for security teams (we are their tool)

Our Values

Technology

Built on: FastAPI (Python) + LSTM Autoencoder with Self-Attention (PyTorch) + Supabase (PostgreSQL) + Solidity on Base Network (Coinbase L2) + React/Next.js Dashboard. 232 tests. 18 API endpoints. 5 smart contracts on Base Sepolia.

Contact