Speech-To-LaTeX

Our platform creates a seamless experience for technical authors to effectively write mathematical papers in LaTeX format - without needing to learn how to code.

Simply record audio and our tool will translate it into LaTeX document, with integrated PDF compilation.

Client-Side GitHub

Directories

• Database Code - Generation of initial MySQL databases
• Latex-PDF-Compile - Compilation of LaTeX docs into PDFs
• api - Python API for handling full-stack integration
• modelInfo - Files related to the model
• scrape_latex - Scraping LaTeX files for training and testing
• speech-to-text-api - API code for speech-to-text model
• text-to-tex-api - API code for text-to-TeX model

Technologies

• SvelteKit (Frontend)
• Vercel (Hosting and deployment)
• HuggingFace (Speech-to-text ML model)
• OpenAI API (Text-to-TeX ML model)
• Google Cloud Platform (Model hosting + cloud infrastructure)
• MySQL (User and file storage)
• ProxySQL (Database load balancing)
• Python + Flask (API for full-stack)
• Google Cloud Storage (LaTeX Training Docs)

Contributors

• Logan Barnhart
• Oscar Carlek
• Claudia Chen
• Tobias Jacobson
• Peter LeCavalier
• Paul Rodriguez