Simple Prompt Optimizer

Turn vague ideas into perfect prompts instantly.
A lightweight, robust, and fast Model Context Protocol (MCP) server designed to optimize prompts using state-of-the-art frameworks like CO-STAR and Chain-of-Thought critique.

🚀 Why use this?

Most "prompt improvers" are slow or hallucinate. This MCP server:

• Uses OpenRouter to access top-tier models (like Llama 3.1 405B or Claude 3.5 Sonnet) for optimization.
• Implements Proven Frameworks: Applies the CO-STAR method (Context, Objective, Style, Tone, Audience, Response) for reliable results.
• Critiques Before rewriting: The deep_optimize tool analyzes your prompt's weaknesses before fixing them.
• Fast: Built on FastMCP for minimal latency.

✨ Features

1. instant_improve

Best for: Daily tasks, quick fixes. Rapidly enhances prompts using structured best practices.

• Inputs: prompt, task_type (coding, writing, etc.), detailed (bool).
• Output: A polished, ready-to-use prompt.

2. deep_optimize

Best for: Complex architecture, critical implementations, debugging. Performs a sophisticated Critique & Refine loop:

1. Analyzes your draft for ambiguity and missing constraints.
2. Plans a strategy to fix it.
3. Rewrites the prompt to perfection.

📦 Installation

Prerequisites

• Python 3.10 or higher
• uv (recommended) or pip

Steps

1. Clone the repository:

   git clone https://github.com/Chumbayoumba/simple-prompt-optimizer.git
   cd simple-prompt-optimizer
   

2. Install dependencies:

   # Using uv (Recommended)
   uv sync
   
   # OR using pip
   pip install .
   

⚙️ Configuration

You need an API key from OpenRouter to power the optimization.

1. Create a .env file in the project root:

   cp .env.example .env
   

2. Add your credentials:

   OPENAI_API_KEY=sk-or-your-key-here...
   OPENAI_BASE_URL=https://openrouter.ai/api/v1
   LLM_MODEL=nousresearch/hermes-3-llama-3.1-405b:free
   

   Note: You can change LLM_MODEL to any model supported by OpenRouter (e.g., anthropic/claude-3.5-sonnet).

💻 Usage with Claude Desktop

To use this MCP server with Claude Desktop, add the following to your claude_desktop_config.json:

Windows

%APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "simple-prompt-optimizer": {
      "command": "uv",
      "args": [
        "run",
        "server.py"
      ],
      "cwd": "C:\\path\\to\\simple-prompt-optimizer",
      "env": {
        "OPENAI_API_KEY": "sk-or-your-key...",
        "LLM_MODEL": "nousresearch/hermes-3-llama-3.1-405b:free"
      }
    }
  }
}

Note: Make sure to replace C:\\path\\to\\... with your actual absolute path.

🛠️ Development

Run the server locally for testing:

uv run server.py
# The server will start and listen on stdio

📄 License

MIT License. Free to use and modify.

---

⭐ If you find this tool useful, please give it a star on GitHub!