Claude Code Remote Control vs HQSSH: Which Should You Use?
An honest comparison of Anthropic's Claude Code Remote Control and HQSSH for driving coding agents from your phone: requirements, supported agents, terminal access, notifications and when each wins.
Updated
Anthropic's Remote Control lets you continue a local Claude Code session from the Claude app or claude.ai/code. HQSSH is a mobile SSH app built for running AI coding agents on your own servers. Both keep the work on your machine and both reach your phone, but they solve different problems.
At a glance
| Claude Code Remote Control | HQSSH | |
|---|---|---|
| What it is | A Claude Code feature that relays a session to Anthropic's apps | A mobile SSH client plus an optional open-source daemon |
| Agents | Claude Code only | Claude Code, OpenAI Codex CLI, OpenCode, Aider and any terminal program |
| Sign-in it needs | claude.ai Pro, Max, Team or Enterprise. Team/Enterprise must be enabled by an owner | Whatever your server already uses |
| API keys, Bedrock, Vertex, LLM gateways | Not supported | Supported (SSH only carries the terminal) |
| Connection path | Your machine → Anthropic API → Claude app | Phone → SSH → your machine |
| Interface | Chat view of the session | The real terminal, with an agent keyboard toolbar |
| Shell access on the same screen | No | Yes: git, logs, tests, other tools |
| Multiple machines | Each machine runs its own Remote Control process | One app for every server you can SSH into |
| Notifications | Push from the Claude app | Terminal bell / OSC 9 alerts, plus ntfy push with the app closed |
| Sending images | Attach in the Claude app | Image button uploads to the server and pastes the path |
| Extra model usage | None beyond the session | None: HQSSH never calls a model |
Where Remote Control is the better choice
- You only use Claude Code and you're on Pro or Max. It's built in. Run
claude remote-control(or/remote-controlin a session) and open the Code tab in the Claude app. Nothing else to install. - You prefer a chat UI to a terminal. Remote Control shows the conversation as messages, which some people find easier to read on a small screen.
- You switch between desk and phone mid-conversation. The terminal session and the app stay in sync, so you can type from either.
Where HQSSH is the better choice
- You use more than one agent. Codex, OpenCode and Aider don't work with Remote Control. HQSSH launches each with one tap and treats them all the same way, including attention notifications. Codex's OSC 9 messages show up as readable alerts.
- You authenticate with an API key, a cloud provider or a gateway. Remote Control requires claude.ai subscription auth and talks to the Anthropic API directly. HQSSH doesn't care how Claude Code is authenticated.
- You want the terminal, not just the chat. Approve a change, then run the tests, check
git diffor tail a log on the same screen. - You manage several machines. A home server, a VPS and a work box are just entries in your systems list, reachable over Tailscale or jump hosts.
- Your traffic must stay on your network. HQSSH is plain SSH from your phone to your machine. The daemon is open source and adds no relay.
Can you use both?
Yes. They don't conflict. A common setup is Remote Control for quick Claude chats on a personal Pro/Max account, and HQSSH for everything else: Codex and other agents, API-key or Bedrock setups, and servers where you need a shell.
Setup in two minutes
Remote Control: update Claude Code, sign in with your claude.ai account, run claude remote-control in your project, and open it in the Claude app.
HQSSH: install the daemon on the machine (optional but recommended), add the machine in the app, and tap LAUNCH CLAUDE.
$ curl -fsSL https://hqssh.com/install | shSee the full walkthrough in How to run Claude Code from your phone.