README.md

metadata

title: PromptTune
emoji: 🐠
colorFrom: indigo
colorTo: green
sdk: gradio
sdk_version: 5.48.0
app_file: app/gradio_interface.py
pinned: false
license: mit
short_description: MLOps for Prompt Engineering and Continuous Improvement.

🚀 Intelligent Prompt Optimizer (IPO-Meta)

This project demonstrates a zero-GPU MLOps pipeline using LLM orchestration to automatically improve the system prompt based on continuous user feedback.

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

🎵 PromptTune

MLOps Toolkit for Interactive Prompt Engineering and Optimization

---

📖 Introduction

promptTune is a modular MLOps toolkit designed for experimenting with, optimizing, and managing LLM prompts. It provides a streamlined interface for rewriting prompts, collecting feedback, and iteratively improving prompt performance—all while maintaining robust, auditable records of prompt changes and user interactions.

---

🚀 Features

🤖 LLM Orchestration & Rewriting: Dynamically leverages a Meta-LLM via the OpenRouter API to transform vague user inputs into highly structured, actionable system prompts, ensuring high-quality responses from the final Task-LLM.

♻️ Continuous Prompt Learning: Implements a zero-GPU, feedback-driven loop where sufficient negative user ratings (Rating: 0) automatically trigger the optimization workflow.

⚙️ MLOps Deployment Pipeline: Uses scheduled GitHub Actions to execute the core Python script, automatically versioning, committing, and deploying the newly refined system prompt configuration back to the main branch.

💾 Versioned Configuration Management: Maintains a single source of truth for the active system prompt (master_prompt.json), ensuring reproducibility and enabling future rollbacks.

💻 Gradio Interface & Data Collection: Provides a simple, Python-native web interface for user interaction and securely logs all raw feedback to inform the next nightly deployment cycle.

📊 Observability Log: Includes a dedicated status file (status_log.txt) that tracks the exact date and time of the last successful prompt deployment, offering a clear audit trail.

---

🚀 Installation

1. Clone the repository:

   git clone https://github.com/your-username/promptTune.git
   cd promptTune
   

2. Set up a Python environment:

   python3 -m venv venv
   source venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Configure environment variables:

   • Create a .env file in the project root and add your OpenAI or compatible API key:

     OPENROUTER_API_KEY=your_api_key_here
     

---

⚡ Usage

1. Run the Gradio Web App

python -m app.gradio_interface

• Interact: Enter prompts, view responses, and provide feedback via the web UI.

2. Optimize Prompts via Script

python scripts/optimize_prompt.py

• This script reviews feedback logs and updates the master prompt for improved results.

3. Project Structure

promptTune/
├── app/
│   ├── __init__.py
│   ├── core_logic.py
│   └── gradio_interface.py
├── data/
│   ├── feedback_log.json
│   └── master_prompt.json
└── scripts/
    └── optimize_prompt.py

---

🤝 Contributing

We welcome contributions! To get started:

1. Fork the repository.
2. Create a branch for your feature or fix (git checkout -b feature-name).
3. Commit your changes.
4. Submit a pull request with a clear description.

Please ensure all code is well-documented and tested.

---

📄 License

This project is licensed under the MIT License.

---

Maintained by Manisankarrr

🔗 GitHub Repo: https://github.com/Manisankarrr/promptTune