A comprehensive STEM platform powered by advanced AI technologies, offering tools for formula search, LaTeX generation, proof assistance, and scientific computing.
- Direct Search: This feature is the application's foundational functionality. When a user types the name of a formula or theorem, the application returns a detailed description, a rendered version of the equation, its usage, and a copyable LaTeX code.
- Multi-step Search: Following the implementation of the Direct Search feature, I recognized that users might not always remember or know the exact names of the equations or theorems they seek. To address this, I developed the Multi-step Search feature, which provides users with a series of options to refine their search within STEM fields, helping them locate the desired information more effectively.
-
The primary challenge lies in managing the LaTeX code response from the Gemini API. I've improved the handling of backslashes and special characters through better cleaning functions, though some edge cases still require attention. For example, certain mathematical notations like summations and integrals sometimes need special handling to render correctly.
-
I successfully resolved the deployment challenge by migrating to Railway, which provides more flexible serverless function execution times. This has significantly improved the reliability of the multi-step search functionality in production.
-
Error Handling and User Experience: I've implemented a more robust error handling system with retry mechanisms and user-friendly error messages. The system now gracefully handles various error scenarios including invalid JSON responses and API timeouts, providing users with clear feedback and a simple "Try Again" option.
-
Advanced AI Integration
- Multiple AI model support beyond Gemini
- Specialized AI models for different STEM fields
- Enhanced accuracy and capabilities
-
Extended STEM Coverage
- Comprehensive resources for all STEM fields
- Field-specific tools and calculators
- Interactive simulations and visualizations
-
Research Tools
- Academic paper integration
- Citation management
- Research collaboration features
-
Educational Features
- Interactive learning modules
- Problem-solving assistance
- Step-by-step explanations
-
Professional Tools
- Advanced computation capabilities
- Data analysis and visualization
- Project management features
- Clone the repository:
git clone https://github.com/Seokhyeon315/latex-generator-ai.git
- Install dependencies by running
pnpm install
. - To use AI features, create an
.env
file on root directory of this project, and add API Key wihGOOGLE_GENERATIVE_AI_API_KEY="YOUR_API_KEY"
. You can refer.env.example
. - To run on local developement server,
pnpm run dev
.