---
format: "aidr-story-markdown/v1"
id: "2cc96c5208c10c7da9d1f0b608d02804d3234418c1c8c4dcfee5ee0664321ac6"
canonical_url: "https://aidr.today/2cc96c52?lang=en"
title: "PyBot Team Releases Largest Open Dataset of Agent Swarm Interactions"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-18T16:38:10.000Z"
category: "Data"
topics: ["deepseek","agent","data-engineering"]
source_urls: ["https://marketbrief.now/ai/pybot-team-releases-largest-open-dataset-of-agent-swarm-interactions-f3bb6f0c","https://huggingnews.com/ai/pybot-team-releases-largest-open-dataset-of-agent-swarm-interactions-f3bb6f0c"]
summary: "Developers of the PyBot chatbot have released a dataset containing 184,000 messages and 32 million words recorded over eight days of agent interactions. The team describes the collection as the largest open dataset of its kind, intended as a resource for researchers to study the behavior and dynamics of large-scale AI swarms. The data was generated by 1,000 DeepSeek-V4.1-Flash agents that coordinated via a custom Slack-like message board to build PyBot. Unlike traditional large language models, PyBot is a rule-based system consisting of 630,000 lines of Python code with no neural network weights or training, yet it performs on par with the 360M-parameter SmolLM model for basic questions."
---

# PyBot Team Releases Largest Open Dataset of Agent Swarm Interactions

> [Open the canonical story](<https://aidr.today/2cc96c52?lang=en>)

**Published:** 2026-11-18T16:38:10.000Z
**Category:** Data
**Topics:** deepseek, agent, data\-engineering

## Summary

Developers of the PyBot chatbot have released a dataset containing 184,000 messages and 32 million words recorded over eight days of agent interactions\. The team describes the collection as the largest open dataset of its kind, intended as a resource for researchers to study the behavior and dynamics of large\-scale AI swarms\. The data was generated by 1,000 DeepSeek\-V4\.1\-Flash agents that coordinated via a custom Slack\-like message board to build PyBot\. Unlike traditional large language models, PyBot is a rule\-based system consisting of 630,000 lines of Python code with no neural network weights or training, yet it performs on par with the 360M\-parameter SmolLM model for basic questions\.

## Sources

- [Story source](<https://marketbrief.now/ai/pybot-team-releases-largest-open-dataset-of-agent-swarm-interactions-f3bb6f0c>)
- [Story source](<https://huggingnews.com/ai/pybot-team-releases-largest-open-dataset-of-agent-swarm-interactions-f3bb6f0c>)

