Now you can even remotely control AI agents using your smartphone.

Recently discovered an open-source project that lets you directly control Claude Code, Cursor, and Gemini CLI running on your local computer from your phone while on the subway or at a café. The principle isn't complicated—it runs an agent server locally and connects via SSH from your phone. But turning this into an out-of-the-box solution genuinely solves some practical problems.


Key highlights:

1. Fully open-source, no subscription traps

2. Faster response than expected, with acceptable latency even on 4G networks

3. Multi-platform support, tested working on both Android and iOS

4. Solid security layer, more trustworthy than TeamViewer alternatives

The accompanying resources are even more surprising—the author compiled 100+ free tutorials covering everything from basic RAG systems to multi-agent collaboration. These aren't just "5-minute quick start" gimmicks but step-by-step guides with complete code.


What worries me most about such tools is "remote functionality for the sake of being remote," but this solution has several practical use cases:

- No need to babysit your computer when debugging long-running AI tasks

- Temporarily modify parameters in your local environment while out

- Smoother demonstrations using phone control (more elegant than screen sharing)

There's more in the GitHub repository:


Of course, there are drawbacks. For instance, entering complex commands on mobile remains awkward—better suited for monitoring than intensive operations. But as an open-source project, it already outperforms many commercial products.

There's a troubling trend in the AI tools space: wrapping simple features in "revolutionary" packaging. This project does the opposite by solidifying fundamental needs, making it refreshingly rare.