Core Technology

PadTai’s infrastructure is built on a quant-native foundation, combining advanced natural language processing, large language models, and proprietary scoring methodologies to deliver actionable insights from the chaos of Crypto Twitter.

Our core technology stack includes:

• NLP + LLM Pipelines for Crypto-Specific Twitter Data

PadTai ingests and processes thousands of X (formerly Twitter) posts per hour using NLP pipelines fine-tuned for crypto slang, trading signals, and narrative shifts. Custom LLM agents are used to detect alpha-rich content, extract named entities (e.g., tokens, influencers), and label sentiment direction.

• Sentiment Scoring Methodology

We apply a proprietary sentiment scoring model that quantifies bullish and bearish intensity at both the tweet and token level. The Net Sentiment Indicator is calibrated based on historical price outcomes, crowd behavior, and narrative alignment—resulting in more accurate detection of inflection points.

• KOL Performance Engine

Every trader and KOL tracked on PadTai is evaluated using our real-time performance engine, which calculates:

• Win Rate: The percentage of price increases following a KOL’s tweet

• Prediction Accuracy: Measures conviction strength by comparing price movement post-mention (1d, 7d, 30d)

• PnL-Weighted Scores: Ranks users based on profitable signal generation over time

This enables transparent, merit-based ranking through our CT Leaderboard and Hall of Fame system.

• Technical Overlays

To supplement social sentiment with price-based context, PadTai includes technical overlays such as:

• Relative Rotation Graphs (RRG) for visualizing token momentum relative to a benchmark

• Momentum Shifts and Price Trends to align crowd sentiment with real market action

• Token Mention Charts to track surging attention across crypto assets