Our Story

Built by developers who wanted to spend less time with agents, and more time with their loved ones.

Agent Approve

The problems we couldn't ignore

If you're a developer managing a team of agents, or a vibe coder with a passionate idea, you probably know this story.

Agent idle time

You step away to make lunch, come back twenty minutes later, and your agent is sitting idle. Instead of a commit, it's waiting for you to approve the first command. Nothing got done.

Destructive commands

It's late, you're tired. You approve a multi-line shell command. Next thing you know, your home directory is being deleted as you rush to hit cancel. You need a way to catch dangerous commands before they execute.

Re-creating policy

You get Cursor dialed in with your allow/deny lists. But then Claude Code asks you to approve them all again. Try out Gemini CLI or Codex, same story. Why can't there be a single policy that works across all of them?

Agent observability

Autonomous agents like OpenClaw or Ralph take action with no human in the loop. Did they stray from the plan, or share sensitive data? You need observability: a clear, real-time view from wherever you are.

Chained to your laptop

The promise of AI is that it frees us to enjoy life more, not shackle us to our computers. Walk the dog, grab a coffee, go for a hike. You shouldn't have to choose between shipping code and living your life.

The idea

“What if I could just approve this command from my watch?”

Imagine you set an agent on a task and go for a walk:

  • A few minutes later, you get a notification to approve a new tool.
  • With a single tap, the agent is unblocked.
  • Later, you get another notification when the agent stops.
  • You expected a commit, but it has a question.
  • No problem. You send a voice reply from your phone.

Your agent stays productive, while you enjoy the fresh air. Now imagine this is happening across all of your agents, using all of your favorite tools.

The unlock

When Cursor, Claude Code, Gemini CLI, and OpenClaw added support for approval hooks, everything clicked. A clean, native way to intercept commands before they execute. No hacks, no wrappers, no new tools to learn. Enterprises adopted hooks for governance. We built this for the solo developer.

The architecture

Hooks in your coding tools communicate through an end-to-end encrypted SaaS relay using websockets and Apple Push Notifications. A centralized policy engine handles most approvals and blocks automatically, so you only get notified when it matters.

The apps

We built a native iOS app with a clean, premium experience, supporting the latest iOS 26 glass effects. And an Apple Watch app so you can truly unblock your agents while out for a run.

Why Agent Approve?

Approve from anywhere

Real push notifications to your iPhone and Apple Watch. Review and approve commands while you're away from your desk. Your agents stay productive, and so do you.

Encrypted and private

All data is encrypted end-to-end with keys unique to your account. You choose what gets stored and for how long. We can't see your data, and we never access your source code.

No DIY required

Delivering reliable push notifications across multiple agents with encryption and privacy controls is a project in itself. We've done that work so you don't have to.

Built by a developer, for developers

This project was born from a real need felt every day. We've launched with iOS and watchOS. If there's interest in Android or a self-hosted option, let us know.