Your AI agents work. You live your life.
Get notified when your agent needs you. Approve with a tap. From anywhere.
First access to the 200 founders spots at €4/month, locked for life.
Your phone gets a notification, you open it, you tap Approve, your agent carries on.
How it works
Install the connector
One command, next to your agent. It's open source and it only talks outward.
Your agent asks, your phone buzzes
When the agent stops for permission, you get a notification with what it wants to do and on which machine.
Approve with a tap
Read it, approve or deny. The agent carries on and you get back to your day.
Security, by construction
MonsterTerminal relays your agent's questions, and it's also a terminal on your phone. When you open a session, the stream passes through our server encrypted and is never kept. Only you and your team, with the permissions you set, can open one.
Open-source connector
The piece that runs on your machine is open source: its code will be published before launch. Read it, audit it, build it yourself.
No open ports
The connector connects outward, like a browser does. Nothing listens on your machine, nothing to expose or firewall.
No keys in our custody
Your SSH keys, API keys and tokens stay where they are. We never store them: the assisted setup uses a key once and deletes it, or you run the one-line command yourself.
No terminal content kept
We keep only a log of who approved what, and when. The output of your terminal, in a session or in an agent's request, is never stored on our side.
Works with the agents you already use
If it runs in a terminal and stops to ask, MonsterTerminal can hand you the question.
Join the waitlist
First access to the 200 founders spots at €4/month, locked for life.
Questions
What is MonsterTerminal?
A small connector that runs next to your coding agent, and an app on your phone that receives its questions and, when you need it, opens a terminal session on that machine. When the agent stops to ask for permission, you get a notification and answer with a tap. The agent keeps the same tools and the same terminal: MonsterTerminal brings the question, and the terminal if you want it, to where you are.
Which agents are supported?
Claude Code, Codex and OpenCode out of the box, plus any terminal process that pauses and waits for a yes. If it runs in a shell, it can go through MonsterTerminal.
Can you read my code or my terminal?
The connector forwards the permission request: which agent, on which machine, what it wants to do. When you open a session from your phone, the terminal stream passes through our server encrypted and is never kept. We keep only a log of who approved what and when, never the output of your terminal.
Do I need to open ports or expose my machine?
No. The connector opens an outbound connection, like a browser does. Nothing listens on your machine and your firewall stays as it is.
What if my phone is offline?
The agent simply waits, exactly as it does today. You can always answer from the terminal: the phone is a second place to say yes, not the only one.
What does the founders price include?
The first 200 paying customers get MonsterTerminal at €4 per month, and that price never goes up for them. The waitlist gives you priority access to those spots, not a guaranteed one. No card is required to join: you pay only when the service launches, and only if you want in.
When does it launch?
A private beta opens first: we pick a small group among the people who ticked the beta box. Everyone on the list hears from us before the public launch.
Is the connector really open source?
Yes. The code that runs on your machine will be published under an open-source license before launch, so you can read exactly what leaves your computer, or build it yourself.
Who can open a session on my machines?
Only you and your team, with the permissions you set for them. Nobody else can start one.