Mindsy.ai

Mindsy.ai is a web application that allows users to upload documents, ask context-aware questions, and receive AI-generated responses. Built with a seamless React frontend and a robust FastAPI backend, the platform is designed to support professionals in domains like finance, legal, and healthcare by simplifying information retrieval from large documents.

Features

1. Secure Document Upload

   • The platform ensures data privacy and secure handling of uploaded files.

2. Context-Aware AI Responses

   • Provides highly accurate and context-aware answers based on the uploaded document.
   • Ideal for professionals in finance, legal, healthcare, and more.

3. Image and OCR Support

   • Extracts text from image-based documents with high accuracy using OCR (Optical Character Recognition).
   • Can read and derive insights from images embedded in docs including graphs, pie charts, etc. making it perfect for financial use-cases.

4. Markdown-Formatted Responses

   • AI-generated responses are displayed in Markdown syntax, ensuring readability with formatted text, headings, bullet points, and code blocks.

5. High-Performance Platform

   • Optimized backend ensures fast query processing with minimal latency.
   • Can handle large documents and complex queries efficiently.

In the above example, the app accurately understands a pie chart from a specifically mentioned page and provides a response in minimal time.

Tools and Technologies Used

• FastAPI: Handles API endpoints for document uploads and query processing.
• Python: Used to implement the backend logic, including integration with LLMs.
• Uvicorn: ASGI server for running the FastAPI backend.
• Gemini 1.5 Flash API: Processes user queries and generates responses using uploaded document context.
• React.js: Component-based library for building the user interface.
• React-Markdown: Displays AI responses with proper Markdown formatting.

Running the Project

You can visit https://mindsy-ai.vercel.app to checkout the interface. To run it locally, follow the steps provided below.

1. Clone the Repository

git clone https://github.com/ashmit0920/mindsy.ai.git  
cd mindsy.ai  

2. Get an API key from Google

1. Head to Google AI Studio and get an API key for the Gemini 1.5 Flash model. It is free of cost and has a rate limit of 15 RPM and 1 million TPM (tokens per minute).

2. Open the sample.env file in the root directory and paste your API key.

   gemini_key = "YOUR_API_KEY" # do not rename the variable 'gemini_key'.

3. Setting Up the Frontend

1. Navigate to the landing-page directory:

   cd landing-page  

2. Install the dependencies:

   npm install  

3. Start the React development server:

   npm start  

   The frontend will run at http://localhost:3000/.

4. Setting Up the Backend

1. Navigate to the root directory:

   cd ..

2. Create a Python virtual environment (optional but recommended):

   python -m venv env  
   env/Scripts/activate  # On Linux, use `source env/bin/activate`  

3. Install the required Python dependencies:

   pip install -r requirements.txt  

4. Start the FastAPI backend using Uvicorn:

   uvicorn main:app --reload  

   The backend will run at http://127.0.0.1:8000/.

4. Testing the Platform

1. Open the frontend in a web browser at http://localhost:3000/.
2. Navigate to the "Get Started" page via the sidebar to access the chat interface.
3. Upload a document using the file uploader.
4. Enter a query and receive AI-generated responses based on the uploaded document.

Folder Structure

mindsy.ai/  
│  
├── landing-page/  
│   ├── src/  
│   │   ├── components/   # React components (Sidebar, Chat Interface, etc.)  
│   │   ├── App.js        # Main React file  
│   │   └── index.js      # Entry point for the React app  
│   └── public/           # Static assets (e.g., GIFs, images)  
│  
├── main.py            # FastAPI application
├── gemini.py          # Gemini API logic
├── app.py             # ignore, deprecated frontend
├── requirements.txt   # Backend dependencies
│  
└── README.md          # Project documentation  

Feedback and Contributions

We welcome feedback and contributions! Feel free to submit issues or pull requests to improve the platform.

---

Thank you for using the Mindsy.ai Platform!