BettaFish

What is BettaFish?

BettaFish (“Micro Opinion” in Chinese) is a program developed by a number ofintelligent body(Multi-Agent) Collaboration-Driven Open Source AIPublic Opinion AnalysisThe tool is designed to break the information cocoon, restore real public opinion and predict trends. Its core innovation lies in a “forum-style” collaboration mechanism that mimics human team discussions, enabling multiple AI intelligences to work together to improve the depth and accuracy of analysis. Launched as a graduation project by developer Guo Hangjiang (screen name BaiFu) in the summer of 2025, the project quickly became a hit on GitHub, accumulating more than 24K starred labels and topping the global trend list.

Development team and background

• Core DevelopersBaiFu is a senior student at the University of Science and Technology of China (USTC), and a representative of “super individuals” in the field of AI.
• technology model: Adopting the “Vibe Coding” development model, combined with theAI Programming Tools(e.g., Claude Code) with human decision-making to achieve fast time-to-market from idea to product.
• Investment support: In March 2026, Chen Tianqiao, founder of Shanda Group, invested 30 million yuan in its advanced version of the product, MiroFish, for technology incubation and commercialization exploration.

BettaFish's main features

1. Network-wide multi-source data collection
   • Covering 30+ domestic and international mainstream social media platforms such as Weibo, Jieyin, Xiaohongshu, Twitter, Reddit and so on.
   • Supports multi-modal content crawling such as text, images, videos, etc., processing millions of data per day.
2. Collaborative multi-intelligence analysis
   • Query Agent: Accurately search domestic and international web pages for information, tracking the source and path of dissemination of events.
   • Media Agent: Parsing multimedia content to extract key information and emotional tendencies.
   • Insight Agent: Mining private databases to correlate with public data for deep insights.
   • Report Agent: Integrate analysis results to generate visual reports (with SWOT analysis, sentiment trend maps, etc.).
3. Trend forecasting and decision support
   • Predict the direction of public opinion based on semantic analysis and heat index modeling.
   • Provide “God's perspective” digital sandbox, support policy preview, crisis public relations simulation and other scenarios.
4. Public-private domain data fusion
   • It supports access to internal databases of enterprises, linking business data with public opinion data for analysis and digging deep insights.
5. Lightweight Deployment and Scalability
   • Pure Python implementation , modular architecture design , support for Docker one-click deployment .
   • Provide API interface for secondary development and integration.

Scenarios for BettaFish

1. Brand PR and Reputation Management
   • Real-time monitoring of public opinion reactions after new product launches and marketing campaigns, identifying negative signs and responding quickly.
   • Case: Brand reputation analysis report of a university, accurately sorting out the time nodes of the incident and public sentiment fluctuations.
2. Media and research organizations
   • Tracking the communication path of hot events, analyzing the law of public opinion evolution, and providing data basis for news reports or policy suggestions.
3. Product Manager & Marketing Analyst
   • Insight into user feedback and demand pain points, optimize product design or marketing strategy to enhance market competitiveness.
4. Education and Research
   • Support research in sociology and communication to explore topics such as public opinion dynamics and emotional computing.
5. Personal Creative Exploration
   • Extrapolate novel endings, simulate entrepreneurial paths, or validate social assumptions to fulfill the need for fun.

How to use BettaFish

• Clone the project and create the environment: Clone the BettaFish project locally from the GitHub repository https://github.com/666ghj/BettaFish and create a new Python environment to install the project dependencies.
• Installation of dependencies: Install the required Python dependency packages in the project directory, and the Playwright browser driver to support the crawler functionality.
• Configuration system: Reproductionconfig.py.examplefile isconfig.py, fill in information such as database configuration and API keys for LLMs (e.g. OpenAI).
• Initializing the database: Run the database initialization script if you choose to use a local database; if you use a cloud database service, you need to contact the project team to apply for access rights.
• activation system: After activating the Python environment, launch the main BettaFish application and access it via a browserhttp://localhost:5000Use the system.

Product Advantages and Future Prospects

• technical barrier: The world's first pure Python implementation of a multi-agent collaborative public opinion system that does not rely on frameworks such as LangChain.
• ecological expansion: The advanced version of MiroFish has been equipped with group intelligent prediction capability, which can cover multiple scenarios such as financial deduction and hotspot simulation.
• social value: Lower the threshold of public opinion analysis, empower individuals and SMEs, and promote the universalization of AI technology.