Last updated: September 14, 2026
LLM Helper is a macOS app that lets you manage AI coding agents (Claude Code, Cursor, Codex, Gemini, and others) running on servers you own or control, by connecting to them over SSH. This policy explains what data the app accesses and what happens to it.
ssh against your servers, it asks you to grant access to your home folder (or
wherever your SSH keys live) once, via macOS's standard security-scoped bookmark permission
dialog. This grant is stored locally by macOS and can be revoked at any time from the app's
Settings tab.LLM Helper connects only to the servers you add, using your own SSH keys, to run commands you initiate (starting a coding agent, checking Docker containers, reading server stats, etc.). All of this happens directly between your Mac and your server — LLM Helper does not operate any backend server of its own, and does not relay your commands, credentials, or server data through any third party.
Some features run install commands for third-party CLI tools (e.g. Docker, GitHub CLI, Node.js, or AI coding agents like Claude Code or Cursor) on servers you specify. Those tools have their own privacy policies, which govern any data you send them once installed. LLM Helper itself does not send your data to those vendors.
All app data (server list, cached statuses) lives locally on your Mac. Uninstalling the app, or deleting its Application Support folder, removes it. Revoking the SSH folder permission (Settings tab → "Grant / Re-select…", or via System Settings → Privacy & Security) removes the app's access to your keys.
If this policy changes, the updated version will be posted at this same URL with a new "last updated" date.
Questions about this policy: info@upg.gr