Do I need to install anything on the phone?

No. termiq.io is a PWA — “Add to Home Screen” and it behaves like an app, including push notifications.

Does it drain the laptop battery or need a public IP?

No port-forwarding, no public IP. The CLI keeps one outbound WebSocket to the relay. Idle overhead is negligible.

The free tier is 1 agent and 2 terminals, no card required. That covers “one laptop, one Claude Code session” completely.

Can several devices watch the same terminal?

Yes — the same session can be open on your desktop and phone; input control follows the device you type from.

Control Claude Code from your phone

Claude Code is great at working alone — until it isn't. It runs for twenty minutes, then stops and asks Allow edit? (y/n). If you've stepped away from your desk, the agent just sits there, waiting. The task you hoped would be done by lunch hasn't moved since you left.

TermIQ fixes exactly this: your terminal follows you to your phone. You get a push notification the moment Claude Code finishes, asks for confirmation, or fails — and you can answer right from the notification, wherever you are.

How it works

Three steps, about two minutes:

Start the agent on your machine
```
npx @termiq/cli
```
The CLI wraps your terminal in a secure relay session. Run claude (or any CLI tool) inside it as usual.
Open termiq.io on your phone and sign in — the web app installs as a PWA, so it feels like a native app with push notifications.
Walk away. When Claude Code stops for any reason, your phone buzzes.

What you get notified about

Event	Notification	What you do
Claude asks y/n or shows a numbered choice	“Terminal is waiting for input”	Tap the answer right in the app
Task finished	“Task completed”	Review the output when convenient
Error / crash	“Terminal error”	Open the live terminal and investigate

The terminal view on your phone is the real terminal — full scrollback, live output, and a keyboard when you need to type an actual command. Permission prompts are parsed into buttons, so answering y/n is one tap, not typing.

Is it secure enough for real work?

This is the part we're strict about, because a terminal stream is about the most sensitive data a developer has:

The terminal stream is end-to-end encrypted between your machine and your phone (XSalsa20-Poly1305). Keys live on your devices.
The server is a blind relay. It routes encrypted bytes; it cannot read your terminal.
The honest caveat: the push notification itself is standard Web Push — not end-to-end encrypted. That's why it only ever carries status metadata (“a terminal needs input”) — not your terminal's own words, unless you turn that on in the agent.

Works with more than Claude Code

TermIQ relays any interactive terminal: Gemini CLI, Codex, aider, plain ssh sessions, long test runs — if it prints output and waits for input, it works. Status detection recognizes the common prompt patterns (y/n, numbered choices, “Press Enter”, plain questions).

FAQ

Do I need to install anything on the phone?: No. termiq.io is a PWA — “Add to Home Screen” and it behaves like an app, including push notifications.
Does it drain the laptop battery or need a public IP?: No port-forwarding, no public IP. The CLI keeps one outbound WebSocket to the relay. Idle overhead is negligible.
What does it cost?: The free tier is 1 agent and 2 terminals, no card required. That covers “one laptop, one Claude Code session” completely.
Can several devices watch the same terminal?: Yes — the same session can be open on your desktop and phone; input control follows the device you type from.

More guides

Stop babysitting your agent.

npx @termiq/cli

Put your phone in your pocket and let Claude Code come to you. Start free — 1 agent, 2 terminals, no card.