Advanced Prompt Generator

The Advanced Prompt Generator is an LLM-based tool designed to automate prompt engineering. It takes a simple input prompt and uses AI agents to enhance it into a highly structured, detailed, and optimized prompt following established engineering principles.

Key Capabilities:

• Automated Enhancement: Transforms minimal user input into complex prompts with specific requirements, structures, and personas.
• Agentic Workflow: Uses LLM agents to generate not just the prompt, but also reference suggestions and a detailed thought process (subtasks, reasoning, and success criteria).
• Deployment Options: Available as a local script, a Gradio web interface, or a scalable FastAPI backend containerized with Docker.

When you provide an input prompt, the generator produces a multi-part response designed to provide maximum context to an LLM. The output typically includes:

1. Enhanced Prompt: A detailed version of your input including a persona (e.g., "As a knowledgeable writing coach"), specific requirements, and a requested response structure.
2. Reference Suggestions: A list of external resources, books, or documentation that would help the LLM fulfill the prompt, including the purpose and how to integrate them.
3. Thought Process: A breakdown of the prompt engineering logic, organized into Subtasks. Each subtask contains:
   • Description: What needs to be done.
   • Reasoning: Why this step is necessary.
   • Success criteria: How to verify the step was completed correctly.

To use the tool on your local machine, clone the repository, install the required Python dependencies, and execute the test script.

Prerequisites:

• Python 3 installed
• OpenAI API Key (the tool supports gpt-4o or gpt-4o-mini)

Steps:

1. Clone the repository.
2. Install dependencies via pip.
3. Run the Advancd_Prompt_Generator.py script to start testing.

git clone https://github.com/Thunderhead-exe/Advanced-Prompt-Generator.git
pip install -r requirements.txt
python3 Advancd_Prompt_Generator.py

The project includes a Gradio web interface for interacting with the prompt enhancement pipeline. When the script is executed, it launches a web UI with the following controls:

• Input Prompt: A multi-line textbox for entering the raw prompt.
• Select Model: A radio selection between gpt-4o-mini (recommended) and gpt-4o.
• Temperature: A slider ranging from 0.0 to 1.0 (recommended: 0.0).

Running the script directly via Python will start the local server.

python Gradio-app/app.py

To use the advanced prompt generation service, you must provide an OpenAI API key. The recommended method is to create a .env file in the project root and add the following key:

OPENAI_API_KEY = your_personal_api_key

Alternatively, you can pass the key directly in pipeline.py, though this is not recommended for security reasons.

OPENAI_API_KEY = the_personal_api_key

The advancedPromptPipeline is an asynchronous function used to enhance a given input prompt using a specified LLM model and temperature setting. It wraps the PromptEnhancer logic to provide an enhanced prompt string.

Parameters:

• InputPrompt (str): The raw text prompt you wish to enhance.
• model (str, optional): The LLM model to use. Supported values are "gpt-4o-mini" and "gpt-4o". Defaults to "gpt-4o-mini".
• temperature (float, optional): The sampling temperature for the model. Defaults to 0.0.

Returns:

• str: The enhanced version of the input prompt.

# Example usage of the pipeline function
advanced_prompt = await advancedPromptPipeline("Your raw prompt here", model="gpt-4o-mini", temperature=0.0)
print(advanced_prompt)

The InputPrompt class defines the schema for the request payload sent to the /advanced_prompt_generation endpoint.

class InputPrompt(BaseModel):
    text: str