Neura Market
BettaFish logo

BettaFish

Free

微舆：人人可用的多Agent舆情分析助手，打破信息茧房，还原舆情原貌，预测未来走向，辅助决策！从0实现，不依赖任何框架。

Text-to-SpeechFreeFree tier
Type
Open Source
Visit Website GitHub (40.7k stars)

About BettaFish

BettaFish (微舆, WeiYu) is an innovative multi-agent public opinion analysis system built from scratch in Python. It orchestrates five specialized components: three analysis agents (QueryEngine, MediaEngine, InsightEngine), one report generation agent (ReportEngine), and a unique coordination layer (ForumEngine) that implements a debate moderator model. The system collects and analyzes public sentiment across 30+ social media platforms (Weibo, XHS, Douyin, Bilibili, etc.) and international news sources, breaking information cocoons and enabling data-driven decision-making. Key innovation is the ForumEngine's chain-of-thought collision, where an LLM moderator analyzes agent communications to identify biases and guide agents, preventing single-model homogenization.

Key Features

Multi-source data collection from 30+ social media platforms and news sources using MindSpider crawler with Playwright automation
Parallel agent execution with three specialized agents (QueryEngine, MediaEngine, InsightEngine) each having distinct LLM backends and tool sets
ForumEngine coordination layer implementing debate moderator model for chain-of-thought collision to reduce bias
Report generation with multi-stage pipeline: template selection, layout, word budget, chapter generation, and multi-format rendering (HTML, PDF, Markdown)
Support for PostgreSQL/MySQL databases and Docker deployment
Built entirely from scratch in Python without dependency on external frameworks

Pros & Cons

Pros
  • Open-source and fully customizable, built from scratch in Python
  • Innovative ForumEngine reduces single-model bias through debate-like coordination
  • Supports 30+ social media platforms and international news sources
  • Parallel agent execution speeds up analysis
  • Includes comprehensive report generation with HTML, PDF, and Markdown output
Cons
  • Requires significant setup: multiple LLM backends, database, and crawler configuration
  • Dependent on external LLM APIs or local models for agent reasoning
  • Documentation is technical and may not be beginner-friendly
  • Latest index date suggests potential maintenance gaps (last indexed May 2026)

Best For

Public opinion monitoring and analysis across multiple social media platformsBreaking information cocoons by synthesizing diverse viewpointsDecision support for enterprises or governments based on sentiment trendsAcademic research on multi-agent coordination and debate moderation

Alternatives to BettaFish

ailearning

AiLearning：数据分析+机器学习实战+线性代数+PyTorch+NLTK+TF2

Awesome-Chinese-LLM

整理开源的中文大语言模型，以规模较小、可私有化部署、训练成本较低的模型为主，包括底座模型，垂直领域微调及应用，数据集与教程等。

CV

✅（已完结）超级全面的 深度学习 笔记【土堆 Pytorch】【李沐 动手学深度学习】【吴恩达 深度学习】【大飞 大模型Agent】

500-AI-Machine-learning-Deep-learning-Computer-vision-NLP-Projects-with-code

500 AI Machine learning Deep learning Computer vision NLP Projects with code

best-of-ml-python

🏆 A ranked list of awesome machine learning Python libraries. Updated weekly.

AI-For-Beginners

12 Weeks, 24 Lessons, AI for All!

FAQ

What platforms does BettaFish support for data collection?
It supports 30+ social media platforms including Weibo, XHS (Xiaohongshu), Douyin, Bilibili, as well as international news sources.
How does the ForumEngine work to reduce bias?
The ForumEngine implements a debate moderator model where an LLM moderator analyzes agent communications via log files, identifies knowledge gaps and biases through synthesize_discussion(), and broadcasts guidance via logs/forum.log. Agents consume this guidance via read_forum_insights().
What are the main components of the system?
The system consists of five components: QueryEngine (web/news research), MediaEngine (multimodal analysis), InsightEngine (private database analysis), ReportEngine (report generation), and ForumEngine (coordination layer).
Is BettaFish free and open-source?
Yes, BettaFish is open-source and hosted on GitHub under the 666ghj/BettaFish repository. Its pricing model is free.