DEV Community

Jsoned
Jsoned

Posted on

AutoWealth AI: An Open-Source Multi-Agent Investment Analysis Engine

#python #ai #finance #opensource

What is AutoWealth AI?

AutoWealth AI is an open-source, multi-agent investment analysis engine built entirely in Python. It uses three specialized AI agents that analyze stocks from different perspectives, then combines their signals through a weighted voting mechanism.

GitHub Repository | 482 tests passing | 11 major features

The Core Architecture

Three AI Agents Working Together

  1. Technical Analyst Agent (35% weight) - Analyzes price action using 12+ technical indicators: MACD, RSI, Bollinger Bands, KDJ, OBV, ATR, DMI, CCI, Williams %R, PSY, and moving average systems.

  2. Fundamental Analyst Agent (35% weight) - Evaluates valuation metrics including P/E ratio, P/B ratio, dividend yield, and growth trends.

  3. Sentiment Analyst Agent (30% weight) - Assesses market sentiment through price momentum, volume patterns, and volatility analysis.

Weighted Voting Decision System

Each agent produces a signal (BUY/SELL/HOLD) with a confidence score. The coordinator aggregates these using configurable weights to produce a final investment decision with target price and stop-loss recommendations.

Key Features

  • Extended Technical Indicators: OBV, ATR, DMI, CCI, Williams %R, PSY
  • Multi-Data Source Support: Yahoo Finance, East Money (akshare), Binance API
  • Machine Learning Prediction: Random Forest + MLP Neural Network, 18-dimensional feature engineering
  • Backtesting System: Strategy backtesting with performance metrics (Sharpe ratio, max drawdown, win rate)
  • Portfolio Optimization: Markowitz mean-variance optimization, efficient frontier
  • Real-Time Alert System: 5 alert types, multi-channel notifications (DingTalk, Feishu, Slack)
  • Social Sentiment Analysis: Twitter/X, Weibo, Reddit, 360+ sentiment words
  • FastAPI Web Service: 7 RESTful endpoints with Pydantic models
  • Chinese NLP Interface: Intent recognition, entity extraction, 80+ stock name mappings
  • Flutter Mobile App: Cross-platform iOS + Android, dark sci-fi theme

Tech Stack

Component Technology
Core Engine Python 3.10+
Data Sources yfinance, akshare, Binance API
ML Models scikit-learn
Web UI Streamlit + Plotly
API FastAPI + Pydantic
Mobile Flutter + fl_chart
Testing pytest (482 tests)

Quick Start

ngit clone https://github.com/Jsoned/autowealth-ai.git
cd autowealth-ai
pip install -r requirements.txt
streamlit run examples/app.py

Project Stats

  • 482 tests all passing
  • 11 major features fully implemented
  • 12+ technical indicators
  • 3 data sources
  • 7 REST API endpoints

AutoWealth AI is designed for educational and research purposes. It is not financial advice.

If you find this project useful, please give it a star on GitHub!

Top comments (0)