Safety and accountability infrastructure for AI agents that act — and pay — on their own.
Two products, one thesis: an autonomous agent needs a gate before the action and a record after it.
| What it does | Try it | |
|---|---|---|
| Black_Wall | Pre-action risk gate for AI agents. Before an irreversible action (send, pay, delete, deploy), one API call returns a risk score, named red flags, and a GO / CAUTION / STOP recommendation with an AUTO / CONFIRM / HUMAN_REQUIRED gate. | npx -y blackwall-mcp · free tier |
| Blackwall x402 oracle | Pre-signature payment-risk verdict (GO / HOLD / STOP) before an agent signs an x402 payment — counterparty reputation, price-anomaly detection, sanctions screening. Verdict only, never custody. Live on Base. | x402.blackwalltier.com |
| Traceipt | Audit-ready receipts for x402 payments. Ed25519-signed the moment settlement confirms on chain, Merkle-anchored, verifies offline against published keys. VAT-aware PDF export. Self-hostable. | request access |
Two locks: Black_Wall gates the payment before it runs; Traceipt proves it after it clears.
-
traceipt-verify— independent verifier for Traceipt receipts:npx traceipt-verify receipt.json(npm). Zero dependencies; proves a receipt without the Traceipt server. -
blackwall-sdk— Python SDK (PyPI) -
blackwall-mcp— MCP server for Claude, Cursor, Windsurf, and any MCP host (npm) -
blackwall-integrations— drop-in guards for LangChain, CrewAI, AutoGen, Vercel AI SDK, Stripe, PayPal, Coinbase, and more -
agent-egress-proxy— localhost egress control for AI agents: observe/enforce every outbound destination, stdlib-only Python
Machine-readable overviews: blackwalltier.com/llms.txt · traceipt.xyz/llms.txt


