SayLO — AI Interview Simulator

SayLO is an open-source interview preparation platform that focuses on privacy and local-first AI processing. It provides simulated interviews, resume parsing, video interviews (via Jitsi), and performance analytics.

Clean, modular architecture with separated frontend and backend.

Quick Links

• Frontend: frontend/src/main.tsx
• Backend: backend/index.js
• Frontend Dev: npm run dev
• Backend Dev: npm run dev:backend
• Both: npm run dev:all

Table of contents

1. Features
2. Getting started (dev & production)
3. Environment variables
4. Testing
5. Project structure
6. Ollama (optional) and AI
7. Contributing
8. Troubleshooting

---

1. Features

• Local-first AI interview simulations (optional Ollama integration)
• Resume parsing (client-side PDF processing)
• Jitsi-based video interviews
• Session storage using IndexedDB / Dexie
• Performance metrics and analytics

2. Getting started

Prerequisites

• Node.js v18+ (Node 20 recommended)
• npm (or pnpm/yarn)

Clone

git clone https://github.com/hitesh-kumar123/saylo.git
cd saylo

Install

npm install

Run locally (two terminals)

• Terminal A — start backend:

npm run dev:backend

• Terminal B — start frontend:

npm run dev

Or run both in one terminal:

npm run dev:all

Open http://localhost:5173 in browser. Backend API runs at http://localhost:3001/api.

Production build

npm run build
npm run preview

Notes

• The repository doesn't run a single combined process by default. Running frontend and backend in separate shells is the simplest local workflow.

3. Environment variables

Create a backend .env at project root and a frontend .env.local in the root for Vite variables.

Backend (.env)

PORT=3001
JWT_SECRET=your_jwt_secret_here
FRONTEND_ORIGIN=http://localhost:5173

Frontend (.env.local)

VITE_API_URL=http://localhost:3001/api
VITE_JITSI_DOMAIN=meet.jit.si
VITE_OLLAMA_HOST=http://localhost:11434
VITE_OLLAMA_MODEL=llama3.2:3b

Tip: Vite only exposes variables prefixed with VITE_ to client code.

4. Testing

Frontend tests with Vitest:

npm test              # Watch mode
npm run test:run      # Single run
npm run test:ui       # UI dashboard
npm run test:coverage # Coverage report

Test files: frontend/test/

Project Structure

saylo/
├── frontend/                  # React + TypeScript frontend
│   ├── src/
│   │   ├── components/        # Reusable UI components
│   │   ├── pages/             # Page components
│   │   ├── services/          # API & utility services
│   │   ├── store/             # Zustand state management
│   │   └── types/             # TypeScript types
│   └── package.json
│
├── backend/                   # Node/Express API
│   ├── index.js               # Main entry point
│   ├── config/                # Configuration files
│   ├── middleware/            # Express middleware
│   ├── routes/                # API endpoints
│   ├── services/              # Business logic
│   ├── utils/                 # Utilities & validators
│   └── db/                    # Database (mock, replace with real)
│
├── docs/                      # Documentation
├── test/                      # Frontend tests
└── package.json               # Root package config

6. Ollama (optional) — local LLM

SayLO can integrate with Ollama to provide richer AI question generation and analysis. Ollama is optional — SayLO has fallback logic when Ollama is not available.

Basic steps (high level):

1. Install Ollama (download from https://ollama.ai or use the platform installer).
2. Run ollama serve (default port 11434).
3. Pull a model, e.g. ollama pull llama3.2:3b.
4. Configure VITE_OLLAMA_HOST and VITE_OLLAMA_MODEL in .env.local.

See OLLAMA_SETUP.md for a more detailed guide and troubleshooting tips.

7. Contributing

Small contribution guidelines are in docs/CONTRIBUTING.md. In short:

• Fork and use feature branches
• Run tests and linters before opening a PR
• Add tests for new functionality
• Use clear, small commits

8. Troubleshooting

• If resume upload / parsing fails: check browser console for PDF.js worker errors and ensure the file is a valid PDF.
• If video (Jitsi) fails to connect: verify browser camera/microphone permissions and that VITE_JITSI_DOMAIN is reachable.
• If AI features are missing: either Ollama isn't running or the configured model isn't available (ollama list).

If you encounter environment-specific issues, open an issue with a short reproduction and logs.

---

If you'd like, I can also:

• Add a docs/ folder with a short Developer Setup and Contributing guide (I already added docs/CONTRIBUTING.md).
• Add a small README badge summary and a short architecture diagram in docs/ARCHITECTURE.md.

Done — updated documentation to be accurate, concise and actionable.