# Experimenting with multiple AI agents collaborating directly on a desktop — retained record

> Retained reddit record from reddit:ollama.

Canonical: https://highsignal.app/data/records/6e8b4aa7bad1bf53

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ollama)  published 2026-11-03 08:02 ingested 2026-11-04 02:41

[original source ↗](https://www.reddit.com/r/ollama/comments/1xmn5q7/experimenting_with_multiple_ai_agents/)

## retained text



```
I've been experimenting with multi-agent collaboration in Bangstick, an AI-first desktop environment I'm developing.

The agents appear as little desktop characters, each with its own role. They can communicate, divide tasks, and work together while interacting with desktop applications.

The system supports ollama, Codex, Claude Code, local LLMs, REA, and Jev.

This video shows the agents working together directly on the desktop.

Still refining the interactions and coordination. Curious what you think of this approach.

  
[https://bangstick.space/](https://bangstick.space/)
```



document fetched 2026-11-04 00:40 · canonical https://reddit.com/r/ollama/comments/1xmn5q7/experimenting_with_multiple_ai_agents
