Quick Start

You will need:

  • A server with SSH (Linux or macOS).

  • An API key from Claude, OpenAI, or any compatible provider or a self-hosted endpoint.

  • A client device — iOS, MacOS, or Linux desktop.

Get the app

Download on the App Store

Download for macOS

Get it on Linux

Get it on Linux

SSH Setup (easy mode)

  1. Click Connect via SSH -> Automated Pairing -> Begin Pairing.
SSH Pairing Dialog
  1. On your server, run:
curl -LsSf https://get.fewshell.com | bash

Enter the 6-digit code.

  1. Verify the information. Use the Test Connection button, if all works, hit save.

NOTE: If you are on a LAN, behind a NAT or VPN, you may need to update the IP address in the dialog.

How it works:

  1. The client generates a public-private SSH key pair.
  2. The public key is pulled by the installation+pairing script using the one-time 6 digit code. (We use a simple service to relay your public key.)
  3. Your new public key is added to ~/.ssh/authorized_keys on your server. Private key is kept on the keychain on the client device.

If you prefer not using our relay, you can manually generate the key pair and paste the private key into the SSH tunnel dialog on the app.

Adding LLM configuration

  1. Tap on the upper-left menu icon. Go to Settings -> Add Model
  2. Configure the settings, test until works, save.
Add AI Model Dialog

Example Workflow

> Check disk usage on all servers

Fewshell will suggest and execute the appropriate commands, presenting results in a readable format.

Next Steps