BruceLanLan/augur

What it solves

Augur addresses the problem of "black-box" AI financial tools that provide vague, unverified answers. Instead of just giving a prediction, it acts as a local research memory system that helps you track what you know, how your understanding has evolved over time, and where different investment experts disagree based on specific evidence.

How it works

Augur functions as a local-first system that runs on your machine, ensuring your data remains private. It simulates the investment philosophies of 18 famous investors (such as Warren Buffett or Ray Dalio) to analyze stocks. It tracks every piece of evidence with its source, business timing, and acquisition time. When data is missing, it explicitly marks it as such rather than guessing, causing the expert models to abstain and explain why. Each research run is saved as an immutable "RunBundle," allowing you to export, compare, and review your research history months later.

Who it’s for

It is designed for serious researchers and investors who need a reproducible, evidence-based workflow rather than quick, unsubstantiated tips. It is particularly useful for those who want to maintain a long-term "thesis journal" and perform rigorous pre-earnings preparation or risk reviews using SEC filings.

Highlights

• Evidence-Based Research: Every conclusion is linked to specific evidence IDs, allowing you to see exactly what facts support or contradict a claim.
• Expert Consensus & Disagreement: Aggregates insights from 18 distinct investment personas, highlighting where they agree and where they clash.
• Local-First & Private: All data is stored locally in ~/.augur with no telemetry, ensuring your research remains yours.
• Reproducible Workflows: Every analysis is saved as a versioned RunBundle, enabling you to audit your past reasoning and compare it against new data.
• Integrated Tools: Includes a dashboard for valuation (DCF), earnings calendars, thesis tracking, and a CLI for managing research cycles.