https://github.com/sayed-ashfaq/stocksnapai
RAG app that helps to analyze the stock, news, kpi's and reports
https://github.com/sayed-ashfaq/stocksnapai
Last synced: 10 months ago
JSON representation
RAG app that helps to analyze the stock, news, kpi's and reports
- Host: GitHub
- URL: https://github.com/sayed-ashfaq/stocksnapai
- Owner: sayed-ashfaq
- Created: 2025-08-10T07:25:44.000Z (12 months ago)
- Default Branch: main
- Last Pushed: 2025-08-10T08:37:58.000Z (12 months ago)
- Last Synced: 2025-08-10T10:14:23.148Z (12 months ago)
- Language: Jupyter Notebook
- Size: 6.84 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
---
# 🤖 AI Finance Copilot (Prototype)
> ⚠️ **Prototype App** – This project is an experimental build to test and validate the idea of an AI-powered finance assistant.
> 🚀 A full **industry-grade web application** is planned for future development.
AI Finance Copilot is a **Streamlit-based prototype** that assists traders and analysts with **stock sentiment analysis** and **financial document insights**. It combines real-time market intelligence with AI-powered document analysis to demonstrate what a next-generation financial research tool could look like.
---
## ✨ Features
### 📈 Stock News & Sentiment Analysis
* Add multiple stocks to a custom portfolio
* Get **real-time news analysis** with AI sentiment scoring
* Portfolio-level insights: themes, risks, opportunities
* Individual stock summaries with sentiment impact
* Portfolio distribution charts (bullish, bearish, neutral)
### 📄 Document Chat & Analysis
* Upload financial documents (earnings, annual reports, etc.)
* AI-powered summaries with **red flag detection**
* Chat with documents using natural language (with citations)
* Quick questions for faster insights
* Export chat history
---
## 📊 Dashboard
#### Link: https://stocksnapai.streamlit.app/
## 🖼️ Screenshots
### Home & API Configuration
### Stock News & Portfolio Analysis
### Document Upload & Summaries
### Interactive Document Q\&A
---
## 🚀 Getting Started
### 1. Clone Repository
```bash
git clone https://github.com/sayed-ashfaq/StockSnapAI
cd StockSnapAI
```
### 2. Install Dependencies
```bash
pip install -r requirements.txt
```
### 3. API Keys (Free Tiers Available)
* [OpenAI](https://platform.openai.com/api-keys) – embeddings & analysis
* [Google AI](https://makersuite.google.com/app/apikey) – Gemini model
* [Tavily](https://app.tavily.com/) – real-time news search
Enter keys in the **sidebar** after starting the app.
### 4. Run the App
```bash
streamlit run main.py
```
---
## 🛠️ Tech Stack
* **Streamlit** – Web interface
* **LangChain + LangGraph** – AI orchestration
* **OpenAI Embeddings** – Document vectorization
* **Google Gemini 2.0** – LLM for chat & analysis
* **Tavily Search** – Real-time financial news
---
## 📂 Project Structure
```
ai-finance-copilot/
├── main.py # Streamlit app entry point
├── requirements.txt # Dependencies
└── modules/
├── stock_analyzer.py # Stock sentiment logic
└── document_chat.py # Document RAG + Q&A
└──...
└──...
```
---
## ⚠️ Prototype Limitations
* Basic UI/UX
* No persistent data storage
* Limited error handling
* Session-based API key storage
* Charts & analytics are minimal
---
## 🔮 Roadmap (Industry-Grade Version)
Planned features for the full-scale application:
* Advanced dashboards & visualizations
* Persistent storage (DB-backed portfolio & documents)
* Real-time alerts & monitoring
* Multi-user support with authentication
* Enhanced security for sensitive documents
* Web based app
* Exception handing and custom logging
* Advanced vectorstore chunking, retriving methods
---
## 📞 Contact
For collaboration or updates on the full version:
* **Email**: [302syedashfaq@gmail.com](mailto:302syedashfaq@gmai.com)
* **LinkedIn**: [sayed-ashfaq](https://www.linkedin.com/in/sayed-ashfaq/)
---
⚖️ **Disclaimer**: This prototype is for demonstration purposes only. It does not provide financial advice.
⭐ Star the repo if you find this project interesting!
---
## Notes
### 📊 Vector DBs Comparison for RAG Apps (2025)
| Feature / DB | **Chroma** 🟢 | **FAISS** 🟠 | **Pinecone** 🔵 | **Weaviate** 🟣 | **Milvus** 🔴 | **Qdrant** 🟤 |
| ----------------------------------- | ----------------------------- | ---------------------------------- | --------------------------------- | ---------------------------- | -------------------------- | --------------------------------------- |
| **Type** | Local DB | Library (ANN) | Managed Cloud DB | Hybrid (self-host + cloud) | Open-source DB | Open-source DB |
| **Persistence** | ✅ Yes | ❌ Manual | ✅ Yes | ✅ Yes | ✅ Yes | ✅ Yes |
| **Metadata filtering** | ✅ Yes | ❌ No | ✅ Yes | ✅ Yes | ✅ Yes | ✅ Yes |
| **Multimodal support** (text+image) | ⚠️ Limited (custom) | ❌ No | ✅ Yes (new multimodal support) | ✅ Yes (via modules) | ✅ Yes (via plugins) | ✅ Yes (via payload) |
| **Scalability** | ⚠️ Small–medium (10k–1M docs) | ✅ Billions (single machine) | ✅ Billions (distributed) | ✅ Billions (distributed) | ✅ Billions (distributed) | ✅ 100M+ (distributed) |
| **Cloud-managed option** | ❌ No | ❌ No | ✅ Yes | ✅ Yes (Weaviate Cloud) | ✅ (Zilliz Cloud) | ✅ (Qdrant Cloud) |
| **Self-hosting** | ✅ Yes (simple) | ✅ Yes | ❌ Cloud only | ✅ Yes | ✅ Yes | ✅ Yes |
| **Ease of use** | ⭐⭐⭐⭐ | ⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐ | ⭐⭐⭐ | ⭐⭐⭐⭐ |
| **Community/Docs** | Medium | High (researchers) | Very High (enterprise) | High | High | Growing fast |
| **Best for** | Prototypes, small RAG apps | Hardcore performance, custom infra | Production SaaS, enterprise scale | Enterprise + semantic search | Open-source, huge datasets | Mid-scale open-source, Rust-based speed |
---
#### ⚡ For Your Case (News + Financial Reports + Images in India)
* Start with **Chroma** → simple, fast iteration.
* If you **need multimodal embeddings (text + image)**: