๐ Inspiration
With the rising popularity of cryptocurrencies, the demand for real-time updates and streamlined insights is higher than ever. We wanted to build an intelligent, modular system that not only fetches live crypto prices but also consolidates breaking news and generates custom alertsโall powered by specialized AI agents.
๐ก What it does
FinCrypt.ai is a modular suite of FastAPI-based AI agents designed for real-time cryptocurrency data aggregation and insights. It features:
- ๐ PriceAgent: Fetches live prices using the CoinGecko API.
- ๐ฐ NewsAgent: Aggregates top crypto news via NewsAPI.
- ๐จ AlertAgent: Simulates alerts based on crypto price thresholds.
- ๐ค AssistantAgent: Acts as a central interpreter and routes user queries to relevant agents.
- ๐ Front End: A simple yet interactive HTML UI for seamless interaction with the AssistantAgent.
๐๏ธ How we built it
- Built each agent using FastAPI for lightweight API services.
- Integrated CoinGecko and NewsAPI for real-time data.
- Used uvicorn for serving each microservice on different ports.
- Designed the AssistantAgent to intelligently parse user input and delegate tasks.
- Created a minimal front end using HTML/CSS/JavaScript, served via a local server to avoid CORS issues.
- Orchestrated the entire flow through RESTful endpoints.
๐ง Challenges we ran into
- ๐ Managing API rate limits and secure key storage (NewsAPI).
- ๐ Ensuring smooth inter-agent communication and query routing.
- ๐ Handling CORS during local front-end testing.
- ๐งต Maintaining modularity while ensuring all components stay in sync.
- ๐ Parsing diverse user inputs accurately in the AssistantAgent.
๐ Accomplishments that we're proud of
- Successfully built and deployed a modular, multi-agent AI system.
- Seamless integration of live data sources (prices + news).
- Developed a clean and interactive UI to demonstrate core functionalities.
- Enabled full conversational querying with multi-agent response coordination.
๐ What we learned
- Deepened our understanding of microservices architecture.
- Gained hands-on experience with FastAPI, API integration, and agent-based design.
- Learned to efficiently manage async API calls and environment variables.
- Practiced modular coding principles for maintainable and scalable applications.
๐ฎ What's next for FinCrypt.ai
- ๐ช Multi-Coin Enhancements: Allow batch querying and watchlists.
- ๐ง Smarter NLP Parsing: Improve AssistantAgentโs ability to understand complex queries.
- ๐ Real Alerts: Trigger actual price notifications based on live thresholds.
- ๐จ UI Overhaul: Rebuild the front end using React or Vue for better UX.
- ๐พ Caching & Rate Limiting: Optimize API usage and performance.
- โ๏ธ Cloud Deployment: Host services via Docker + cloud platforms for global accessibility.
Built With
- agentverse
- fetchai
- nextjs
- python
Log in or sign up for Devpost to join the conversation.