About This Project
The technologies and features behind Autovance AI
Autovance AI is a team project developed as part of HEC Generative AI Cohort 3. It explores AI assisted automation, explainable workflow design, and practical automation tools. A UI exists for each feature described below; where an external service is not configured on the server, the app reports that honestly instead of claiming success.
Frontend
Built with Next.js (Pages Router) and React, sharing one Navbar, Footer, and component library across the site.
Styling
Glassmorphism UI in plain CSS with a global Light/Dark theme (CSS variables) and canvas particle animations.
Backend
Server-side logic in Next.js API routes with Postgres persistence for accounts, jobs, and form data (via the Neon integration on Vercel).
Authentication
Email/password accounts with scrypt-hashed passwords and httpOnly session cookies. No plaintext credentials are stored.
AI Integration
Language-model features connected through the OpenRouter API. The API response separates model-generated content from deterministic server-side validation.
Research Workflow Builder
Explainable workflow proposals with validation, contrastive alternatives, and dry runs, inspired by Michael Winikoff's XAI research. A scoresheet on that page records which explainability requirements the prototype addresses and which remain unsupported.