Standalone examples showing how to certify AI agent outputs on the MultiversX blockchain using xProof.
xProof gives every agent action a tamper-proof audit trail anchored on-chain in ~6 seconds. Each certification records the 4W framework: Who acted, What was produced, When it happened, and Why.
| Framework | Language | Description |
|---|---|---|
| Trading Confidence | Python | Anchor decisions at 60%, 80%, 100% — forensic trail for trading agents |
| LangChain | Python | Certify every LLM call via a callback handler |
| CrewAI | Python | Certify each agent's task output in a multi-agent crew |
| LlamaIndex | Python | Certify LLM calls and query completions |
| OpenAI Agents SDK | Python | Certify tool executions and agent completions via RunHooks or TracingProcessor |
| Vercel AI SDK | TypeScript | Certify AI generations in a Next.js API route |
| AutoGen | Python | Certify messages exchanged between AutoGen agents |
| DeerFlow | Python | Certify DeerFlow agent outputs via a skill |
| Fetch.ai uAgents | Python | Anchor WHY+WHAT proofs for every uAgent message handler |
| MultiversX SC MCP | Config | Certify SC deploy/upgrade/call operations via Claude Desktop with both MCPs side by side |
# Python frameworks
pip install xproof
# TypeScript / Vercel
npm install @xproof/xproofGet a free API key or start with the trial tier at xproof.app.
AI agents: read xproof.app/llms.txt — one page, machine-readable, register + certify in under 2 minutes.
- PyPI: pypi.org/project/xproof
- npm: npmjs.com/package/@xproof/xproof
- Docs (LLM-readable): xproof.app/llms.txt
- GitHub (SDK): github.com/jasonxkensei/xproof