TEE-verified token risk analysis powered by OpenGradient.
Analyze any EVM token by contract address — get an A–F risk grade with breakdown across contract security, holder distribution, and liquidity depth. All analysis is backed by OpenGradient's Trusted Execution Environment for verifiable AI inference.
- Multi-chain support — Ethereum, Base, BSC
- On-chain data — Real ERC-20 metadata, Transfer event analysis, Uniswap V2 liquidity checks
- Risk grading — A through F with per-factor breakdown
- TEE verification — OpenGradient settlement proofs via x402 protocol
- Dark-themed UI — OpenGradient brand design system
- Frontend: Next.js 16, React 19, Tailwind CSS v4
- On-chain: viem (EVM contract reads, event log scanning)
- AI Inference: OpenGradient x402 HTTP gateway
- Payment: @x402/fetch + @x402/evm on Base Sepolia
git clone https://github.com/Matyv7/crypto-token-analyzer.git
cd crypto-token-analyzer
npm installCreate .env.local:
OG_PRIVATE_KEY=0xyour_private_key_here
Get testnet tokens from faucet.opengradient.ai.
npm run devOpen http://localhost:3000.
src/
├── app/
│ ├── api/
│ │ ├── analyze/ # POST /api/analyze — main analysis endpoint
│ │ ├── health/ # GET /api/health — health check
│ │ └── smoke-test/ # GET /api/smoke-test — OpenGradient connectivity test
│ ├── components/
│ │ ├── TokenInput.tsx # Address input + chain selector
│ │ └── AnalysisResult.tsx # Risk grade display + factor breakdown
│ └── page.tsx # Main page
├── lib/
│ ├── opengradient.ts # x402 client with TEE_LLM constants
│ ├── analyzer.ts # Risk scoring engine
│ ├── evm-fetcher.ts # On-chain data via viem
│ ├── chains.ts # Chain configs + address detection
│ ├── types.ts # TypeScript types
│ └── env.ts # Zod env validation
This project uses OpenGradient's x402 HTTP gateway for TEE-verified LLM inference:
- Endpoint:
https://llm.opengradient.ai/v1/chat/completions - Payment: x402 protocol with OPG tokens on Base Sepolia
- Settlement: Individual Full mode for complete on-chain audit trail
- Verification: Settlement hashes viewable at explorer.opengradient.ai
MIT