Android AppFunctions 🤝 on-device MCP You can create self-describing functions using the AppFunctions Jetpack library to let assistants like Gemini execute tasks via natural language. Connect your app to AI agents with Android AppFunctions →

This looks promising, but without access to Gemini it’s hard to meaningfully evaluate or demo AppFeatures. For teams trying to plan product roadmaps, the ability to test real integrations is critical. Any chance to open early access or a developer preview so we can validate?

does this work cross-app? like can Gemini call functions registered by a totally different app?

we don't care, we don't want your ai crap, people want open Android

@backlon More power with AI 👨‍💻🌍

on-device MCP via Jetpack AppFunctions is such a clean bridge. if Gemini can call these reliably offline, thats a big shift for app automation.

We think so!

It look promising! Actually I never seen like this kind of approach for applications.

The best part is that these AppFunctions execute locally on the device, providing a native, secure way for Gemini to interact with your app.

Hello @androiddev, does Android support IEEE 802.11af and IEEE 802.11ah? @grok, which Wi-Fi, Bluetooth or other communication protocols are supported by Android?

This is how mobile apps become part of the agent execution layer, not just interfaces.

MCP + self-describing functions is way more interesting than another “AI-native” wrapper.