This project demonstrates how to extract structured data from PDFs using Google's Gemini 2.0 AI model in a Next.js web application. It allows users to upload PDFs and dynamically generate JSON schemas based on user prompts, which are then used to extract structured information from the documents.
How It Works:
- Upload PDF: Users can upload their PDF documents through the web interface
- Define Schema: Users provide a natural language prompt describing the data they want to extract
- Schema Generation: Gemini 2.0 generates a JSON schema based on the user's prompt
- Data Extraction: The Schema is used to extract structured data from the PDF using structured output from Gemini 2.0
- Results: Extracted data is presented in a clean, organized format
- 📄 PDF file upload and preview
- 🤖 Dynamic JSON schema generation using Gemini 2.0
- 🔍 Structured Outputs using Gemini 2.0
- ⚡ Next.js frontend with shadcn/ui
- 🎨 Uses Gemini 2.0 Javascript SDK
First, set up your environment variables:
cp .env.example .envAdd your Google AI Studio API key to the .env file:
GEMINI_API_KEY=your_google_api_key
Then, install dependencies and run the development server:
npm install
npm run devOpen http://localhost:3000 with your browser to see the application.
- Build the Docker image:
docker build -t pdf-structured-data .- Run the container with your Google API key:
docker run -p 3000:3000 -e GEMINI_API_KEY=your_google_api_key pdf-structured-dataOr using an environment file:
# Run container with env file
docker run -p 3000:3000 --env-file .env pdf-structured-dataOpen http://localhost:3000 with your browser to see the application.
- Next.js - React framework for the web application
- Google Gemini 2.0 - AI model for schema generation and data extraction
- shadcn/ui - Re-usable components built using Radix UI and Tailwind CSS
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.