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

Description and use

The author built a 17‑agent AI swarm on an Android phone using Termux and Python. Each agent performs a single focused task, and agents communicate via a file‑based message queue. This enables local automated data analysis and report generation, though it is resource‑intensive and technically complex.

Problem

Running a distributed, multi‑agent AI system entirely on a mobile device with limited resources.

Outcome

The author successfully deployed a 17‑agent swarm on an Android phone, achieving decentralized intelligence for automated data analysis and report generation, albeit with limited efficiency and high technical complexity.

Workflow

• Install Termux from F‑Droid to get a Linux environment on the phone.
• Set up a lightweight Python environment and install dependencies (requests, sqlite3).
• Create SQLite database for data storage.
• Write individual agent scripts (e.g., data acquisition, preprocessing, analysis, report generation) each performing a single focused task.
• Implement a file‑based message queue: each agent publishes JSON messages to target agents’ input directories and subscribes by polling its own input directory.
• Use threading to run multiple agents concurrently within the same Python process or separate processes.
• Run agents in Termux, monitor logs, and iterate on error handling and performance.
• Validate that agents collaborate to produce aggregated reports and alerts.

Evidence

I wanted a network of smaller, specialized agents. Each agent would be relatively simple, but collectively they could tackle tasks beyond the capacity of any single agent.

Source

The biggest constraint was, of course, the limited resources of a phone. Forget heavy frameworks like TensorFlow or PyTorch. I needed something lightweight and efficient.

Source

I ended up with a surprisingly effective solution: a file-based message queue.

Source