Last updated
Last updated
StrawberryAI's Twitter Agent continuously monitors socials to detect key opinion leaders' (KOLs) plugs and mentions, analyzing engagement across follower networks, and autonomously identify emerging narratives. By processing data in real-time, the platform provides always up-to-date scores and highlights on new (and established) projects and protocols, delivering actionable intelligence directly from quality, processed data.
The Twitter Agent is designed to monitor a curated list of Key Opinion Leaders (KOLs), ranked meticulously by their influence and spam activity. Every minute, a robust data ingestion pipeline scans for new tweets related to each KOL, ensuring that no significant sentiment shift goes unnoticed. When new tweets are identified, they trigger a series of data classification pipelines that operate seamlessly in the background:
Sentiment Analysis: Each tweet is assigned a sentiment score ranging from 0 (extremely bearish) to 10 (extremely bullish). This nuanced scoring system allows for a precise understanding of market sentiment in real time.
Ticker Detection: Tickers are identified through explicit mentions or inferred from context, images, and previous discussions. This capability enables the system to capture emerging trends and narratives that might otherwise go unnoticed.
Further Classification: Detected tickers undergo additional extraction and classification processes, ensuring that each piece of information is contextualized within the broader market landscape.
For any previously unseen ticker, a Researcher Agent conducts thorough investigations across reputable platforms like CoinGecko, DexScreener, and CoinMarketCap to gather essential information:
Project Insights: Detailed descriptions and key points are saved for future reference and updates, providing a rich database of knowledge that can be tapped into at any time.
Classification Pipelines: Projects are categorized into specific sectors (e.g., Ethereum Ecosystem, Memecoin, Lending Protocol), allowing for targeted analysis and strategy development.
Macrocategory Classification: A broader narrative classification (e.g., DeFi, Meme, AI) is applied, ensuring that users can easily identify overarching trends and themes.
Embedding Pipeline: Extracted data is transformed into embedding tokens, facilitating easy retrieval by AI agents, akin to a "Google search" for cryptocurrency. This transformation significantly enhances the system's efficiency and effectiveness in delivering relevant insights.
