uOttahack8 - Token Talk

TokenTalk was created for uOttahack8, hosted at the University of Ottawa (Winter 2026). Project uses OpenRouter SDK to evaluate the efficiency of several Large Language Models when processing a user task. The results are evaluated objectively utilzing modern machine learning processes.

Techstack:

Front-end: React with Typescript, styled with Tailwind CSS Back-end: Python and Node.js Hosting: Github pages and AWS Lambda API: OpenRouter

Project Structure

src/
├── App.tsx            # Main React app entry; renders EntryPage and sets up global styles/background
├── App.css            # App-wide CSS, including gradient backgrounds and scoped styles
├── api/
│   └── translate.ts   # API utility for sending translation requests to the backend
├── assets/            # Static assets like images and icons
│   └── react.svg      # Example asset
├── components/        # React components organized by feature or page
│   ├── EntryPage.tsx  # Landing page with animated background; launches the multi-step wizard
│   ├── LandingPage.tsx # Multi-step wizard for role/language selection, file/text input, and validation
│   └── ResultsPage.tsx # Displays translation results; allows copying or restarting
├── index.css          # Tailwind CSS imports and base styles
├── main.tsx           # React root renderer; mounts App to the DOM

server/
├── index.js           # Node.js Express backend for the translation API
├── .env               # Environment variables (e.g., API keys, model config) – ignored by git
└── package.json       # Backend dependencies and scripts

Local Set Up Instructions

1. Create the .env file in src/server with the following:

OPENROUTER_API_KEY=<YOUR_OPENROUTER_API_KEY>
OPENROUTER_MODEL=openai/gpt-4o-mini
// Other models

2. In one terminal, run npm run dev
3. Navigate to src/server, and run npm install, then node index.js