I built a 17-agent AI swarm on my phone — here's how

Description and use

A 17-agent AI swarm was built to run entirely on a smartphone using Pyodide and PWA, enabling local data processing and decision making without cloud dependence.

Problem

Running a distributed, resilient, and private AI swarm entirely on a smartphone without cloud dependencies.

Outcome

The swarm achieved real-time sensor monitoring, activity recognition, and adaptive device settings entirely on-device, demonstrating edge AI feasibility.

Workflow

• Wrap Python code in an HTML page with Pyodide
• Install the page as a PWA for native-like experience
• Define 17 agents (sensors, processors, decision makers, communicators) as Python classes
• Use asyncio event loop to run agents concurrently
• Implement inter-agent communication via a Python queue
• Process sensor data, make decisions, and trigger actions through browser APIs
• Iterate continuously in a feedback loop

Evidence

I wrapped my Python code in a simple HTML page using Pyodide and then used a progressive web app (PWA) approach to "install" it on my phone.

Source

Each agent is a Python class responsible for a specific task. My swarm consists of 17 agents, broadly categorized as: Sensors, Processors, Decision Makers, Communicators.

Source

The biggest hurdle was, unsurprisingly, execution. Standard Python interpreters on Android (like QPython) are often limited and can feel sluggish. That’s where Pyodide came in.

Source