Dexter: An Autonomous Agent for Deep Financial Research

Introduction

Dexter is an autonomous financial research agent designed to think, plan, and learn as it works. It performs in-depth analysis by utilizing task planning, self-reflection, and real-time market data. Envision it as a specialized tool, akin to Claude Code, but specifically engineered for comprehensive financial research.

It takes complex financial questions and transforms them into clear, step-by-step research plans. Dexter then executes these tasks using live market data, critically checks its own work, and refines the results to provide confident, data-backed answers.

Important Disclaimer: This project is intended for educational, entertainment, and informational purposes only. It is not designed for real trading or investment activities. Users should be aware that outputs may be incorrect, incomplete, or outdated, and it does not offer financial, investment, tax, or legal advice. Always consult a licensed financial advisor before making investment decisions.

Installation

To get Dexter up and running, follow these steps:

Prerequisites

You will need:

• Bun runtime (v1.0 or higher)
• OpenAI API key
• Financial Datasets API key
• Exa API key (optional, for web search)

Installing Bun:

For macOS/Linux:

curl -fsSL https://bun.com/install | bash

For Windows:

powershell -c "irm bun.sh/install.ps1|iex"

Verify installation:

bun --version

Setup Steps

1. Clone the repository:

   git clone https://github.com/virattt/dexter.git
   cd dexter
   

2. Install dependencies with Bun:

   bun install
   

3. Set up your environment variables:

   cp env.example .env
   # Edit .env and add your API keys
   # OPENAI_API_KEY=your-openai-api-key
   # FINANCIAL_DATASETS_API_KEY=your-financial-datasets-api-key
   # EXASEARCH_API_KEY=your-exa-api-key
   

Examples

Dexter offers various ways to interact and evaluate its capabilities.

How to Run

Run Dexter in interactive mode:

bun start

For development with watch mode:

bun dev

How to Evaluate

Dexter includes an evaluation suite to test its performance against financial questions.

bun run src/evals/run.ts

To run on a random sample:

bun run src/evals/run.ts --sample 10

Results are logged to LangSmith for detailed analysis.

How to Use with WhatsApp

Integrate Dexter with WhatsApp to chat and receive responses directly:

bun run gateway:login
bun run gateway

Then, message yourself on WhatsApp with a question for Dexter.

Why Use Dexter

Dexter stands out with its advanced capabilities designed for sophisticated financial analysis:

• Intelligent Task Planning: Automatically breaks down complex queries into structured research steps.
• Autonomous Execution: Selects and utilizes the appropriate tools to gather financial data efficiently.
• Self-Validation: Critically checks its own work and iterates on tasks until completion, ensuring accuracy.
• Real-Time Financial Data: Provides access to up-to-date income statements, balance sheets, and cash flow statements.
• Safety Features: Includes built-in loop detection and step limits to prevent unintended runaway execution.

These features make Dexter a powerful, self-improving agent for anyone looking to conduct deep financial research with an AI-driven approach.

Links

• GitHub Repository: virattt/dexter
• Follow on Twitter: virattt
• Join Discord Server: Dexter Discord