-
Live tweet tracking
Inspiration
Vaccine acceptance around the world has been hindered by misinformation. We want to solve that by delivering factual information from reliable sources to the right people at the right time.
What it does
Use a cohort of autonomous agents to:
- Track the web for comments, messages, posts about vaccination.
- Analyse these inputs with LLMs, to feed a live dashboard for human monitoring
- Suggest Action plan to deliver the right information to the people mentionning misinformation, fear, or other unjustified/misguided negative sentiment towards vaccination
- Search for realtime reliable sources with PerplexityAI
- Track user engagement and evolution of views after taking action
- Iteratively refine action plans for different user profiles based outcomes of previously taken actions.
How we built it
Use FetchAI uagents to build the agents cohort PerplexityAI for web search OpenAI/Anthropic/Google LLMs for data analysis, decision making and message drafting, with careful prompt engineering Tweeter, reddit, facebook, email etc, for interacting with Online users.
Challenges we ran into
Data collection and live deployement
Accomplishments that we're proud of
Successfully build an interactive dashboard and an agentic workflow
What we learned
How to use and deploy uAgents from FetchAI
What's next for !Plague
Attaching a real web-scrapping engine and engagement tracking module
Log in or sign up for Devpost to join the conversation.