Autonomous financial research agent that transforms complex queries into structured, data-backed answers. It's designed for technical users needing deep financial analysis, offering capabilities akin to advanced LLM code assistants but specialized for market data. The agent plans, executes, validates, and refines its research autonomously.

How It Works

Dexter employs a multi-agent architecture comprising Planning, Action, Validation, and Answer agents. Complex financial questions are first decomposed into structured research tasks by the Planning Agent. The Action Agent then selects and executes appropriate tools to gather real-time financial data (income statements, balance sheets, cash flow). Crucially, the Validation Agent verifies task completion and data sufficiency, allowing the agent to iterate and refine its work until a confident, data-backed answer is synthesized by the Answer Agent. This approach ensures robust analysis and prevents premature or inaccurate results.

Quick Start & Requirements

• Primary install/run command: uv sync for dependencies, uv run dexter-agent for interactive mode.
• Prerequisites: Python 3.10+, uv package manager, OpenAI API key, Financial Datasets API key (from financialdatasets.ai).
• Setup: Clone the repository, sync dependencies with uv, copy env.example to .env, and populate with API keys.
• Links: Repository: https://github.com/virattt/dexter.git

Highlighted Details

• Intelligent Task Planning: Automatically decomposes complex queries into structured research steps.
• Autonomous Execution: Selects and executes the right tools to gather financial data.
• Self-Validation: Checks its own work and iterates until tasks are complete.
• Real-Time Financial Data: Access to income statements, balance sheets, and cash flow statements.
• Safety Features: Built-in loop detection and step limits to prevent runaway execution.

Maintenance & Community

The repository includes contribution guidelines but does not specify community channels (e.g., Discord, Slack) or notable maintainers/sponsors.

Licensing & Compatibility

This project is licensed under the MIT License, which is permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

The agent's execution is governed by configurable safety limits (max_steps, max_steps_per_task) to prevent indefinite loops or excessive resource consumption, indicating a need for careful configuration.