🔥 Prompt Plugin

AI prompt enhancement toolkit that turns vague prompts into structured, high-quality instructions for better results from ChatGPT, Claude, Gemini, and other AI chatbots.

The Problem

Most people type quick, unstructured prompts into AI chatbots and get generic, shallow responses. Prompt engineering makes a huge difference in output quality — but writing detailed, well-structured prompts every time is tedious and requires expertise most users don't have.

The Solution

Prompt Plugin sits between you and any AI chatbot. You type your rough idea, and it automatically:

1. Detects what you're trying to do — coding, writing, analysis, debugging, brainstorming, etc.
2. Adds expert role framing — e.g. "You are an expert software engineer" for coding tasks
3. Applies tone styling — matches your chosen tone (professional, casual, technical, etc.)
4. Boosts specificity — if your prompt is too vague, it adds detail and example requests
5. Injects structure — asks the AI to organize its response (sections, bullet points, code + explanation)
6. Adds quality guardrails — instructs the AI to prioritize accuracy, flag uncertainty, and use examples

The result is a prompt that consistently produces better, more useful AI responses — without you having to think about prompt engineering.

What It Does

❌  "sort a list in python"

✅  "You are an expert software engineer. Write Python code that sorts a list.
    Follow best practices, include comments, handle edge cases, and provide
    example usage."

❌  "explain docker"

✅  "You are a patient teacher who explains complex topics clearly.
    Explain the concept of Docker. Include a simple overview, a real-world
    analogy, key terminology, a practical example, and common misconceptions.
    Be warm, approachable, and encouraging."

Who It's For

• Developers who use AI for coding help and want better code output
• Writers & marketers who need AI-generated content that actually sounds good
• Students & researchers who want clear, structured explanations
• Teams who want a shared set of prompt templates and consistent AI interactions
• Anyone who uses ChatGPT, Claude, or Gemini daily and wants better results with less effort

How It Works

Prompt Plugin uses a rule-based enhancement pipeline (no API keys needed, no external AI calls). Everything runs locally:

• Category Detection — keyword analysis across 9 task categories to classify your prompt
• Template Engine — 12 pre-built templates with variable substitution for common tasks
• Enhancement Pipeline — 6-stage processing chain: role framing → tone → specificity → structure → context → guardrails
• History Manager — local JSON-backed storage for saving, searching, and starring past prompts

Available as four interfaces:

• Web UI — browser-based dashboard with live enhancement
• REST API — FastAPI with Swagger docs, integrate into any app
• CLI — terminal tool with rich formatting, clipboard copy
• Python library — import and use directly in your code

Features

• Auto-Enhance — rewrites vague prompts with role framing, specificity boosts, and guardrails
• 12 Templates — coding, writing, analysis, brainstorming, debugging, summarization, translation, explanation
• 6 Tones — professional, casual, technical, creative, academic, friendly
• Context Injection — append extra context and system instructions
• Prompt History — save, search, star, and reuse past prompts
• Auto Category Detection — detects coding, writing, analysis, etc. from prompt text

Install

pip install -e ".[dev]"

Usage

CLI:

prompt-plugin enhance "write a python function to sort a list"
prompt-plugin enhance "explain docker" --tone friendly --context "audience is beginners"
prompt-plugin templates
prompt-plugin history

Web UI + API:

prompt-plugin serve
# Web UI → http://127.0.0.1:8000
# API docs → http://127.0.0.1:8000/docs

Python:

from prompt_plugin.engine import PromptEngine
from prompt_plugin.models import PromptRequest, Tone

engine = PromptEngine()
result = engine.process(PromptRequest(
    raw_prompt="write a python function to sort a list",
    tone=Tone.technical,
))
print(result.enhanced)

Deploy

Render (free): Push to GitHub → render.com → New Web Service → connect repo → Deploy.

Docker:

docker build -t prompt-plugin .
docker run -p 8000:8000 prompt-plugin

Tests

pytest -v

License

MIT