BettaFish brings agent swarms to sentiment analysis

This is less a toy “multi-agent app” and more a full vertical stack for social listening and research ops. The real story is that BettaFish ships the messy systems work most agent repos avoid, which is why developers are paying attention.

• –It spans 10+ domestic and international social platforms, comment-level crawling, and multimodal inputs, pushing it beyond a thin LLM wrapper
• –The architecture mixes agents with fine-tuned and statistical sentiment models, which is a stronger design than relying on one model for everything
• –Its ReportEngine, IR layer, and HTML/PDF output pipeline make it useful as an end-to-end research product, not just a chatbot interface
• –Pure Python modularity, Docker support, and explicit extension points make it a strong reference project for teams building domain-specific agent systems