
https://www.loom.com/share/03d30efcf8e44035af03cbfebf840c73
Start Claude Code in the terminal, then pick it up instantly on web or mobile with a seamless handoff between all three
Founded by Kartik Sarangmath & Ishaan Sehgal
Childhood friends since Montessori. They have spent years working in AI and AI infra at big tech and acquired startups and love building tools that remove friction for developers.
Claude Code is incredible, but it chains you to your desk.
Once an agent starts working, you’re stuck waiting in the terminal for it to ask you something. That could take 5–10 minutes, and if you’re away, progress halts. They have coded from their beds, on walks, in Ubers, while doing laundry… and yes, even on the toilet — but only after they built Omnara.
Existing solutions don’t solve this well:
Omnara is the first agent command center — start an AI agent in the terminal, then continue the same session from their web dashboard or mobile app.
With Omnara, you can:

The team built a CLI wrapper that mirrors the native Claude Code experience — output, permissions, notifications, and mode switching — by parsing the session file at ~/.claude/projects and terminal output in real time. All agent and user messages are streamed via SSE to the Omnara platform, where you can respond from anywhere.
It’s fully open source: github.com/omnara-ai/omnara

Omnara works with any AI agent, but Claude Code users immediately “get” the value. It’s the perfect human-in-the-loop use case — and the one they use every day.
They have also used Omnara as a human-in-the-loop node in n8n workflows, e.g., for replying to emails. This is just the beginning.

Free for up to 10 agent sessions/month. $9/month for unlimited.
Then log in at omnara.com or download their iOS app: Omnara on the App Store