AI Feedback Systems

A Flask web app for AI-powered student script marking. Teachers create classes, upload class lists, and mark assignments using Claude, GPT, or Qwen. Supports single and bulk marking, student self-submission, department mode with HOD oversight, and a demo showcase mode.

Quick Deploy

Setup

1. Install dependencies

pip install -r requirements.txt

Requires Python 3.10+ and poppler for PDF-to-image conversion.

2. Set environment variables

Variable	Required	Description
ANTHROPIC_API_KEY	At least one provider key	Anthropic API key for Claude models (console.anthropic.com)
OPENAI_API_KEY	At least one provider key	OpenAI API key for GPT models (platform.openai.com)
QWEN_API_KEY	At least one provider key	Qwen API key for Qwen models
FLASK_SECRET_KEY	Yes (production)	Session secret + encryption key. Generate with python -c "import secrets; print(secrets.token_hex(32))"
APP_TITLE	No	App title shown in nav, pages, PDFs (default: AI Feedback Systems)
TEACHER_CODE	No	Master access code for normal mode. Permanent key that always works.
DEPT_MODE	No	Set to TRUE for department mode with HOD/teacher roles
DEMO_MODE	No	Set to TRUE for demo mode (3 budget models, no auth)
DATABASE_URL	No	PostgreSQL connection URL (default: SQLite marking.db)
PORT	No	Server port (default: 5000)

3. Run

# Development
python app.py

# Production
gunicorn -w 1 --threads 100 --timeout 300 --bind 0.0.0.0:$PORT app:app

Operating Modes

Normal Mode

One teacher per deployment. Set TEACHER_CODE as your access code.

1. Log in with your code
2. Create a class and upload a class list (CSV)
3. Create assignments (upload question paper, answer key)
4. Mark students individually or in bulk
5. Share assignment links for student self-submission

Department Mode (DEPT_MODE=TRUE)

HOD manages multiple teachers and classes.

• HOD creates classes, assigns teachers, manages access codes
• Each teacher sees only their assigned classes
• HOD dashboard with insights, analytics, CSV export
• HOD can revoke or purge teacher accounts
• TEACHER_CODE env var is the HOD master key

Demo Mode (DEMO_MODE=TRUE)

For showcasing the app without real data.

• No authentication required
• Single-script marking with 3 budget models (Haiku, GPT Mini, Qwen 3.5)
• Class/assignment creation is session-only (not persisted)

Demo + Department Mode (DEMO_MODE=TRUE + DEPT_MODE=TRUE)

Pre-seeded HOD dashboard showcase.

• Auto-login as demo HOD
• 3 classes, 3 teachers, 6 assignments with realistic student results
• Full insights and analytics from seed data

Class List Format

CSV with two columns:

Index,Name
01,Alice Tan
02,Bob Lee
03,Charlie Ng

Header row is auto-detected and skipped.

Bulk Marking

Upload a single PDF containing all student scripts. Set the number of pages per student (can vary per student). Set page count to 0 to skip a student and keep their existing result.

Tech Stack

• Backend: Flask, SQLAlchemy, Gunicorn
• AI Providers: Anthropic Claude, OpenAI GPT, Alibaba Qwen
• PDF: ReportLab (generation), pypdf (splitting), pdf2image (conversion)
• Frontend: Vanilla JS, MathJax (LaTeX rendering), Animate.css
• Database: PostgreSQL (production) or SQLite (development)