Documentation · Getting Started

From zero to your first agent in minutes

reading time: 4 minplatforms: macOS · Linux · WSL2license: MIT

1. Install the agent

remoteagent is a small open-source daemon that runs on your own computer. It talks to the control plane at remoteagent.online, where your reasoning happens — your machine stays the only place your files live.

curl -fsSL https://remoteagent.online/agent-installer | bash

The installer places a remoteagent command on your PATH and persists it in your shell profile. Requirements: Node.js 20+, macOS / Linux / WSL2.

2. Generate your Remote Agent key

In the console, open Settings → Remote Agent key and press Generate new. This is the one secret your device needs. The key is shown once — treat it like a password.

Keep your key local. The console never stores it in plain text; it is scoped to your account and can be revoked instantly from Settings → Devices.

3. Connect a device

remoteagent login

Paste the key when asked. The daemon dials out to the control plane — no ports to open, no inbound connections, works behind NAT. Your device appears in the console with live telemetry: CPU, memory, load, disk and uptime.

4. Create your first agent

Back in the console, press + New Agent. Give it a name (e.g. Hermes), pick a device, and optionally a system prompt. The agent starts stopped; start it and open a chat.

Model routing is automatic. No provider key on your machine ever. Simple tasks go to cheap models, complex tasks to deeper ones — the brain config in Settings tunes this per account.

5. Watch it work

Every task streams to the console: reasoning steps, tool calls, tool results, tokens and cost. Open Deep Dive to inspect a full run trace, or Logs for the live event feed. You can watch the agent think and act in real time — nothing runs in the dark.

Optional: enable two-factor authentication

Open Settings → Security & two-factor authentication and scan the QR code with Google Authenticator, Authy or 1Password. With 2FA on, sensitive actions — adding API keys, generating device keys, forgetting devices, factory reset — require a fresh verification code. See Security Model.