IronClaw β Deploy AI agents securely in encrypted enclaves
IronClaw is a secure, open-source platform for running AI agents inside Trusted Execution Environments on NEAR AI Cloud. It stores API keys and passwords in an encrypted vault, injecting them only at approved network endpoints, so the LLM never sees raw secrets. Each tool runs in an isolated WebAssembly sandbox with capability-based permissions, network allowlists, and leak detection. Built in Rust and deployable in one click, it supports OpenClaw-style workflows with stronger memory safety and control.