Hi all,

We’ve been thinking about a core limitation in current mobile AI assistants:

Most systems (e.g., Apple Intelligence, Google Assistant–style integrations) rely on predefined schemas and coordinated APIs. Apps must explicitly implement the assistant’s specification. This limits extensibility and makes the ecosystem tightly controlled.

On the other hand, GUI-based agents (e.g., AppAgent, AutoDroid, droidrun) rely on screenshots + accessibility, which gives broad power but weak capability boundaries.

So we built Mobile-MCP, an Android-native realization of the Model Context Protocol (MCP) using the Intent framework.

The key idea:

- Apps declare MCP-style capabilities (with natural-language descriptions) in their manifest.

- An LLM-based assistant can autonomously discover all exposed capabilities on-device via the PackageManager.

- The LLM selects which API to call and generates parameters based on natural language description.

- Invocation happens through standard Android service binding / Intents.

Unlike Apple/Android-style coordinated integrations:

- No predefined action domains.

- No centralized schema per assistant.

- No per-assistant custom integration required.

- Tools can be dynamically added and evolve independently.

The assistant doesn’t need prior knowledge of specific apps — it discovers and reasons over capabilities at runtime.

We’ve built a working prototype + released the spec and demo:

GitHub: https://github.com/system-pclub/mobile-mcp

Spec: https://github.com/system-pclub/mobile-mcp/blob/main/spec/mobile-mcp_spec_v1.md

Demo: https://www.youtube.com/watch?v=Bc2LG3sR1NY&feature=youtu.be

Paper: https://github.com/system-pclub/mobile-mcp/blob/main/paper/mobile_mcp.pdf

Curious what people think:

Is OS-native capability broadcasting + LLM reasoning a more scalable path than fixed assistant schemas or GUI automation?

Would love feedback from folks working on mobile agents, security, MCP tooling, or Android system design.

Build your own AI Agents in 4 Mins

We do need to get money moving differently

PostgreSQL Statistics: Why queries run slow

Race Engineer for Daily Commute

The complete Manic Miner disassembly

Show HN: A CAPTCHA based on 3D spatial recognition

OpenClaw Partners with VirusTotal for Skill Security

Scarnon

Sudo-Rs Breaks Sudo Tradition of Invisible Password Typing

Guardclaw – 7 layers between your AI agents and disaster

Show HN: What to Watch – aggregated ratings across streaming services

Let OpenClaw be your shipping manager

Happy Map

Block lays off 40% of its staff because of AI, CEO says most CO will do the same

MailVault – Free, open-source email client that stores emails locally as .eml

Mobile-MCP: Letting LLMs autonomously discover Android app capabilities

Jack Dorseys Block to cut nearly half its workforce in AI overhaul, shares surge

Why the Storytellers Will Win the AI Agent Race

OCTP – A cryptographic trust protocol for AI-era open source contributions

Gzpeek: Tool to Parse Gzip Metadata

Creating momentum when an employee is stuck

Accidentally disabling SSH access via scp

Mothlamp Problems

The Economics of Technical Speaking

Exposing More Parallelism Is the Reason Why Some Vectorized Loops Are Faster

AI for Social Impact: Ideas Worth Building

Neanderthals seemed to have a thing for modern human women

LastSaaS: Free, open-source SaaS boilerplate; Go+React, built with Claude Code

Anthropic says it 'cannot in good conscience' allow Pentagon to remove AI checks

DevToolbox – 150 developer tools, no login, no tracking

Show HN: PipelineIQ – AI diagnoses CI/CD failures and sends fixes to Slack