#723 · Primary category: AI Agents & Automation

PhoneClaw

agent-framework ai-agent gemma ios litert local-agent local-ai mobile-agent mobile-agent-framework mobile-ai mobile-ai-agent-framework ollama on-device-agent on-device-ai phone-agent-harness phone-agent-loop phone-ai-agent phone-harness phone-loop swift

PhoneClaw turns phones into local AI agent runtimes with on-device models, native mobile Skills, LiveLand, and optional Mac Gateway inference.

Project last updated:08/06/26

GitHub Stars

1.2K

Forks

166

Contributors

2

License

Apache-2.0

Why we included this project

If you want a private AI assistant that runs entirely on your iPhone, PhoneClaw is worth a look. It executes Gemma and MiniCPM-V models on the device itself via LiteRT, so ordinary requests never leave the phone, and the native Skill system lets the agent read Calendar and Reminders, pull Health data, handle contacts, understand images from the camera, and act from the Dynamic Island through LiveLand, asking for confirmation before risky actions like deletes. An optional Mac Gateway client pairs with a Mac over your local network for heavier inference through Ollama when a task needs more compute. For developers, the framework routes skills and chains tool calls, which makes multi-step requests like health summaries or schedule analysis reasonably reliable, and you can add custom Skills of your own.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category