Senti

SentiTrade is a NEAR-powered AI agent that supports memecoin traders by analyzing Twitter sentiment in real-time to generate buy, sell, or hold recommendations, enhancing trading strategies.

Comment

Inspiration

Volatile memecoin markets need better, decentralized tools for gauging public sentiment. We wanted to give NEAR users an edge with AI-powered insights.

What it does

SentiTrade is a NEAR AI agent that analyzes Twitter for memecoin hype, providing a hype score, sentiment summary, and basic trading recommendation (Buy/Sell/Hold), along with the current Bitcoin price.

How we built it

We built it using NEAR AI (with the llama-3p1-8b-instruct LLM), Python, tweepy for Twitter integration, and vaderSentiment for analysis. The agent utilizes NEAR's built-in tools.

Challenges we ran into

Integrating with Twitter's API and ensuring reliable tweet data fetching proved challenging. We also had to learn how to correctly use the NEAR AI Environment object and tweepy client.

Accomplishments that we're proud of

We successfully built a functional NEAR AI agent that analyzes Twitter sentiment, generates trading recommendations, and fetches real-time Bitcoin price data. We overcame initial challenges with API integration.

What we learned

We learned a lot about NEAR AI agent development, Twitter API integration, sentiment analysis techniques, and asynchronous programming in Python.

What's next for Senti

Future plans include adding support for more memecoins, refining the trading logic, integrating with a NEAR smart contract for on-chain data storage, and exploring cross-chain capabilities.

Built With

Share this project:

Updates