BruceLanLan/augur

What it solves

Augur addresses the difficulty of obtaining diverse, expert-level investment perspectives on a single stock. Instead of relying on one source, it simulates a "investment decision committee" of 18 legendary investors (such as Warren Buffett, Ray Dalio, and Cathie Wood) to provide a weighted consensus, risk-adjusted position sizing, and structured debates on specific tickers.

How it works

The system uses AI agents modeled after the philosophies of 18 famous investors across four schools of thought (Value, Growth, Macro, and Chinese Value). It integrates real-time financial data from sources like SEC EDGAR and yfinance to perform analysis. Users can trigger a multi-step workflow (fetch $\rightarrow$ analyze $\rightarrow$ consensus $\rightarrow$ committee $\rightarrow$ debate $\rightarrow$ sentiment) to get a comprehensive verdict. The project also implements the Model Context Protocol (MCP), allowing AI agents like Claude or Hermes to directly interact with the Augur workspace and tools.

Who it’s for

It is designed for investors and researchers who want to simulate professional investment committee deliberations, compare different investment philosophies, or integrate expert-driven stock analysis into their own AI agent workflows.

Highlights

• Investment Committee Simulation: Pre-set committees (e.g., Classic Value, Macro All-Weather) or custom combinations of legendary investors.
• Quantitative Consensus: Provides an Augur Score (0–10), BUY/NEUTRAL/SELL signals, and Kelly Criterion-based position suggestions.
• Structured Debate: Automatically generates bullish and bearish arguments by pitting selected investors against each other.
• Agentic Integration: Full MCP support with 13 tools, allowing external AI agents to read/modify terminal configurations and run analysis pipelines.
• Real Data Integration: Uses official SEC EDGAR 10-K/10-Q filings for fundamental metrics and tracks insider trading and institutional holdings.