- 🐕 What is PET?
- 🏗️ Visual Architecture
- ✨ Key Features
- 📁 Project Structure
- 🛠️ Installation Guide
- 🚀 How to Use
- 👋 About the Builder
- 📜 License
- 🤝 Contributing & Support
PET (Prompt Enhancement Tool) is a high‑performance browser extension that lives inside your favorite AI platforms. It transforms vague, one‑line queries into 40+ line strategic prompt architectures using professional engineering techniques like Chain‑of‑Thought, Socratic Method, and Few‑Shot prompting.
🎯 Stop chatting with AI. Start architecting with your personal companion.
- 🚀 10x better responses — structured prompts unlock reasoning depth
- 🧠 Learn prompt engineering by example — see how pros build prompts
- ⚡ Zero context switching — works inside your existing chat interface
flowchart TB
User((👤 User)) -->|vague prompt| Ext[🐕 PET Browser Extension]
Ext -->|send to| BE[⚡ FastAPI Backend]
Ext -.->|fallback| Rule[📜 Local Rule Engine]
BE -->|LangChain| AI[🧠 Groq / OpenAI / DeepSeek]
AI -->|strategic prompt| Ext
Ext -->|inject| LLM[💬 Host Chat Interface]
LLM -->|response| Score[📊 Scoring Engine]
Score -->|feedback| User
| Feature | Description |
|---|---|
| 🐾 Smart Rewrite | Turns "explain physics" into a structured, role‑based, multi‑step roadmap. |
| 📊 Response Scoring | Real‑time 0–100 evaluation of AI responses based on goal coverage + reasoning depth. |
| 🧪 Multi‑Model Mastery | Custom‑tuned logic for ChatGPT, Claude, Gemini, and DeepSeek. |
| 🛡️ Fail‑Safe Design | 8‑second timeouts + local rule‑based fallback — works even if backend is offline. |
| 🎨 Premium UI | Glassmorphic sidebar, fluid Lottie animations, zero‑footprint draggable widget. |
| 📈 Growth Dashboard | Track prompt quality trends and model performance over time. |
.
├── 📂 backend/ # FastAPI + LangChain Intelligence
│ ├── main.py # API routes & webhook handling
│ ├── agents.py # Prompt architect logic & decision trees
│ └── memory_store.py # Tab‑isolated session management
│
├── 📂 src/ # Chrome extension source
│ ├── 📂 content/ # UI components (React + Vite)
│ ├── 📂 background/ # Service worker & message bridge
│ └── 📂 engines/ # API logic, scoring, & metrics
│
├── 📂 website/ # Landing page — “The Forest”
│ └── index.html # High‑fidelity marketing site
│
└── 📂 dist/ # Production‑ready extension build
cd backend
pip install -r requirements.txt
# Add your GROQ_API_KEY to .env
uvicorn main:app --reload💡 Optional: Also supports OpenAI / DeepSeek keys — see
agents.pyfor configuration.
- Download the
dist/folder from this repo. - Open Chrome →
chrome://extensions. - Enable Developer Mode (top‑right).
- Click Load unpacked → select the
dist/folder. - 🎉 Your Pet appears in the corner on ChatGPT, Claude, Gemini, or DeepSeek!
- Type a rough idea in the chat box
(e.g.,"how to build a SaaS") - Click the Pet to open the architect panel.
- Hit ▶ Rewrite — PET generates 3 strategic versions of your prompt.
- Use Prompt to inject the version you like into the chat box.
- Score the AI’s reply to see how well it followed the strategy.
🏆 Pro tip: Higher scores (80+) mean the AI deeply followed your prompt architecture. Low scores? Try another rewrite version!
Krishianjan Lanka
Data Grad · AI Engineer · Builder focused on Design & Execution
“I built PET because I got tired of seeing people get mediocre AI responses from vague prompts. PET makes prompt engineering accessible to everyone — from Doctors to Plumbers — allowing anyone to prompt like a professional engineer.”
This project is licensed under the MIT License — see the LICENSE file for details.
- 🐛 Found a bug? Open an issue
- 💡 Have an idea? Pull requests are welcome — we’re
first‑timers‑friendly - 💬 Join the community on Discord (placeholder, replace with your invite)
- ⭐ Star this repo — it helps more people find PET and levels up prompt engineering for everyone
🌿 Built with ❤️ for better conversations with AI.