I built a mobile command center for Claude Code and other AI agents – Omnara

Ever started Claude Code on a task, stepped away for coffee, and come back to find it waiting for your approval? Or had your laptop die mid-session and lost everything?

I kept running into this. I'd kick off a refactoring job before a meeting, but then I'm stuck either babysitting the terminal or missing critical approval requests. The agent would just pause. Waiting. While I'm three floors away.

The solution: agents that run independently

So I built Omnara. It runs AI coding agents as headless processes that stream everything to your phone and browser in real time. The entire agent loop lives on your machine or in the cloud, not trapped in your terminal. You can monitor logs, review diffs, answer questions, and approve next steps from literally anywhere.

Voice coding while you commute

The mobile part isn't just monitoring either. There's a full conversational voice mode. I've used it walking to the subway, talking through architecture decisions with the agent, having it spawn parallel sub-agents for different parts of a feature, all while I'm nowhere near my laptop. It feels like pair programming, but you're commuting.

Built for complex, multi-agent workflows

Sessions are worktree-isolated, so when you have multiple agents running in parallel, changes stay scoped and reviewable. And here's the part that actually changed how I work: seamless handoff between local and cloud. Laptop battery dies? Session keeps running. Switch from your desk to your phone? Full context, chat history, uncommitted code, all there.

Who this is for

I built this because I was tired of being chained to my desk while agents worked. If you've ever felt stuck waiting for long-running tasks, or wished you could code while away from your computer, I'd love to hear if this resonates.

What else should something like this handle?