Aleksey Matsarski
revert README to original config
5224a4e

A newer version of the Gradio SDK is available: 6.5.1

Upgrade
metadata
title: Financial Analysis System
emoji: 💬
colorFrom: yellow
colorTo: purple
sdk: gradio
sdk_version: 5.42.0
app_file: app.py
pinned: false
hf_oauth: true
hf_oauth_scopes:
  - inference-api
license: mit
short_description: Financial analysis system

Multi-Agent Financial Analysis System

Overview

The Multi-Agent Financial Analysis System is an advanced AI-driven platform designed to revolutionize the way financial data is processed and analyzed. The system employs a multi-agent architecture powered by agentic AI, leveraging specialized Large Language Models (LLMs) to analyze market news, earnings reports, and stock data. The ultimate goal is to deliver structured, actionable, and explainable investment insights for traders, investors, and financial analysts.

Objectives

  1. Multi-Agent Architecture
    Design a robust multi-agent system with distinct specialized roles, including:

    • News Analysis
    • Earnings Report Analysis
    • Market Trends Analysis
  2. Data Integration
    Integrate real-time financial data sources such as APIs and news feeds to ensure accurate and timely analysis.

  3. Explainable Insights
    Generate transparent and explainable investment briefs and predictions, enabling users to make informed decisions.

  4. Autonomous Workflows
    Demonstrate autonomous capabilities by showcasing features like:

    • Planning investment tasks
    • Self-critique and evaluations
    • Iterative workflow improvements

Key Features

  • Agent-Based Processing: The system utilizes multiple specialized agents that work collaboratively on segmented tasks to provide a holistic analysis.
  • Real-Time Data Analysis: Connects to live financial data sources and processes high volumes of data in real-time.
  • Actionable Output: Provides structured summaries and final investment recommendations tailored to specific investment scenarios.
  • Transparency: Investment insights are backed by clear explanations, improving trust and usability.

Technology Stack

The system is built using modern AI technologies centered around Large Language Models (LLMs) and integrates with financial data APIs. The architecture includes:

  • Python-based agent workflows.
  • Specialized modules for financial data ingestion and processing.
  • Autonomous feedback mechanisms for iterative improvement and decision-making.

Example Workflow

Here is an example of how the system functions:

  1. A user provides a company stock ticker (e.g., AAPL).
  2. The system begins processing by splitting tasks across agents to analyze:
    • News relevant to the stock.
    • The company’s latest earnings information.
    • Broader market trends impacting the stock.
  3. The agents collaborate, refine insights, and self-critique results, leading to an actionable investment recommendation.
  4. The final analysis provides:
    • News Summary: Key relevant headlines and sentiment analysis.
    • Earnings Summary: Insights from the latest earnings reports.
    • Market Summary: Broader trends and their relationship to the selected stock.
    • Final Recommendation: A structured, actionable output.

How to Get Started

To contribute or explore the system in action:

  1. Clone the project repository.
  2. Configure API keys and data sources as necessary.
  3. Run the main workflow function to generate insights.
from workflow.agents_workflow import build_agents_workflow

app = build_agents_workflow(llm_model_name="gpt-4o-mini")

# Example usage:
state = app.invoke({
    "ticker": "AAPL",
    "query": "Produce investor-ready insights for AAPL",
    "news_summary": None,
    "earnings_summary": None,
    "market_summary": None,
})

Roadmap

Future enhancements include:

  • Expanding the agent roles to support alternative asset classes (e.g., cryptocurrencies, bonds).
  • Integrating more complex sentiment and risk analysis tools.
  • Implementing an intuitive web-based user interface for broader accessibility.

Contributions

We welcome contributions from the community! Please review the contribution guidelines and feel free to submit pull requests or issues with suggestions and improvements.