✨ PromptmaX — AI Prompt Engineering Platform

Turn rough, simple prompts into precision-engineered, production-ready instructions.

Explore Platform • Quick Start • Core Features • Tech Stack

---

PromptmaX is a next-generation AI prompt engineering platform. It uses a sophisticated 9-stage optimization pipeline to dismantle rough, vague user inputs and reconstruct them into highly structured, context-rich, and boundary-safe instructions.

Whether you are building custom AI agents, fine-tuning model behaviors, or running automated prompt pipelines, PromptmaX ensures your AI interactions are deterministic, safe, and consistently high-quality.

• Python 3.10+
• Supabase Account (for cloud database storage - optional)
• Official Mistral AI API Key (for prompt enhancements and scoring)

1. Clone the Repository

   git clone https://github.com/Santosh-Prasad-Verma/PromptmaX.git
   cd PromptmaX

2. Set Up a Virtual Environment

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Configure Environment Variables Copy the example environment file and fill in your keys:

   cp .env.example .env

   Required variable to set:

   • MISTRAL_API_KEY (used for the core prompt engineering models)

   Optional sync variables:

   • SUPABASE_URL, SUPABASE_ANON_KEY, and SUPABASE_SERVICE_ROLE_KEY

4. Initialize the Database & Run

   cd backend
   python manage.py migrate
   python manage.py runserver

   Open http://localhost:8000 in your browser to start using the app!

• ⚡ 9-Stage Enhancement Pipeline: Automates the CREATE prompt algorithm (Context, Role, Exclusions, Action, Theme, Evaluation) with iterative self-healing loops to polish prompts until they hit target scores.
• 📈 6-Dimension Quality Scoring: Real-time quality evaluation analyzing Clarity, Specificity, Structure, Context, Constraints, and Output Formatting with interactive heatmap feedback.
• 🔄 A/B Testing & Variations: Generates three distinct style variations (Concise, Detailed, and Structured) for comparing responses.
• 🌐 Web Search Context Injection: Automatically crawls and pulls real-time information from the web to ground prompts with fresh factual context.
• ** Automated Cloud History Sync**: Safe dual-write database synchronization that auto-saves your prompt enhancements directly to the cloud.

flowchart TD
    Input[📥 Raw User Prompt] --> Stage1[1. Input Validation]
    Stage1 --> Stage2[2. NLP Quality Analysis]
    Stage2 --> Stage3[3. Persona & Context Injection]
    Stage3 --> Stage4[4. CREATE Engine Optimization]
    Stage4 --> Stage5[5. Quality Gating & Validation]
    Stage5 --> Output[📤 High-Fidelity Prompt]

• Frontend: Vanilla CSS (Premium cream & evergreen theme), Vanilla JavaScript, Lucide Icons, Google Fonts (Inter, EB Garamond)
• Backend: Django (Python 3.10+), Django REST Framework
• Database: SQLite (Local development) / PostgreSQL (Production)
• Authentication: Supabase Auth (SSO Integration with JWT validation)
• AI Models: Mistral AI API (exclusive official models)
• Background Tasks: Celery, Redis (for async website scraping and prompt history synchronization)

├── backend/                  # Django backend application
│   ├── enhancer/             # Core prompt logic, views, models, and tasks
│   └── promptx_project/      # Main Django server configuration and settings
├── frontend/                 # Premium responsive frontend client
│   ├── pages/                # Clean public pages and interactive chat workspace
│   ├── scripts/              # Auth bridge, Supabase client, and chat controller
│   └── styles/               # CSS styling sheets (cream theme & premium popups)
├── supabase/                 # Database migrations and dashboard email templates
└── render.yaml               # Infrastructure-as-code deployment blueprint