HackerAI lets you link your own machine or network to the web app through Agent Mode. After you connect, you can run commands on your device from HackerAI—useful for tooling you’ve installed locally, automation, and work against assets on your LAN.
This capability is available on all paid plans.
What the local client does
The HackerAI Local Sandbox Client (@hackerai/local) opens a secure connection between your computer and HackerAI. While it’s running, Agent Mode can:
Execute terminal commands on your computer in real time
Use your machine’s network visibility for scanning and testing (subject to your OS and permissions)
Commands run directly on your host OS—there is no container isolation layer in the current client.
1. Get Your Token
Log in to HackerAI.
Open Settings (from the sidebar or your account menu).
Open the Agents tab.
Click Generate Token or copy an existing token.
That token links this machine to your HackerAI account.
2. Install and Start the Local Agent
Quick Start (recommended)
npx @hackerai/local@latest --token YOUR_TOKEN
Or install globally:
npm install -g @hackerai/local
hackerai-local --token YOUR_TOKEN
3. Usage Examples
Basic Docker Mode (default)
npx @hackerai/local@latest --token hsb_abc123 --name "My Laptop"
--nameis optional; if omitted, the client defaults to your hostname.
4. Options
Option | Description |
--token TOKEN | Required authentication token |
--name NAME | Name shown in HackerAI |
--help, -h | Show help message |
Security Overview
Commands run on your OS without isolation. Only run the client on machines you trust and control.
The client exits automatically after about 1 hour of inactivity (no commands), to reduce long-lived sessions.