Ultimate Prompt Generator (UPG) 🤖

A powerful CLI tool for generating, managing, and reusing prompts for different LLM providers. Currently supports OpenAI and Anthropic models with smart defaults.

Join Russian Speaking Telegram Channel | Watch Russian Video Tutorial

🌟 Features

• 🔄 Multi-Provider Support:
  • OpenAI (default: gpt-4o)
  • Anthropic (default: claude-3-sonnet-20240229)
• 📝 Smart Prompt Generation:
  • Generate high-quality prompts for various tasks
  • Create prompts with or without variables
  • Automatic variable detection and validation
• 💾 Prompt Management:
  • Save and reuse generated prompts
  • Tag and categorize prompts
  • Search through saved prompts
• 🔑 Secure Configuration: Safe storage of API keys and preferences

🚀 Quick Start

# Install with pip
pip install upg-cli

# Or install with Poetry
poetry add upg-cli

Initial Configuration

# Configure your preferred LLM provider
upg config --provider openai --api-key YOUR_API_KEY

Basic Usage

# Generate a simple prompt without variables
upg generate "Write a story about a space traveler"

# Generate a prompt with variables
upg generate "Create a Python function to calculate fibonacci numbers" \
    -v FUNCTION_NAME \
    -v ARGS \
    --save --name "python-fibonacci" \
    --tag python --tag math

# Use a saved prompt
upg answer python-fibonacci \
    -v FUNCTION_NAME "fibonacci" \
    -v ARGS "n: int"

📋 Detailed Usage Guide

Prompt Generation

You can generate prompts both with and without variables:

# Simple prompt without variables
upg generate "Write a poem about autumn"

# Prompt with variables
upg generate "Write a {GENRE} story about {TOPIC}" \
    -v GENRE \
    -v TOPIC

# Save generated prompt
upg generate "Your task description" [options]

Options:
  --provider TEXT           LLM provider to use (openai/anthropic)
  -v, --variable TEXT      Variable names for the prompt (optional)
  -o, --output FILENAME    Save prompt to file
  -s, --save              Save prompt for later use
  --name TEXT             Name for saved prompt
  -d, --description TEXT  Description for saved prompt
  -t, --tag TEXT         Tags for categorizing the prompt

Using Prompts

For prompts without variables, you can use them directly:

# Using a simple prompt without variables
upg answer simple-story

# Using a prompt with variables
upg answer story-template \
    -v GENRE "mystery" \
    -v TOPIC "lost artifact"

More Examples

Simple Prompts (No Variables)

# Generate a blog post outline
upg generate "Create an outline for a blog post about machine learning basics" \
    --save --name "blog-outline" \
    --tag content --tag blog

# Generate coding guidelines
upg generate "Write Python code style guidelines for a team" \
    --save --name "python-guidelines" \
    --tag python --tag guidelines

# Use saved prompts
upg answer blog-outline
upg answer python-guidelines

Prompts with Variables

# Generate a template for API documentation
upg generate "Write documentation for a REST API endpoint" \
    -v ENDPOINT \
    -v METHOD \
    --save --name "api-docs" \
    --tag api

# Use the template
upg answer api-docs \
    -v ENDPOINT "/users" \
    -v METHOD "POST"

🔧 Configuration

UPG provides flexible configuration options through the config command group:

Provider Configuration

# Configure a provider
upg config provider --provider openai --api-key YOUR_API_KEY --model gpt-4o

# Or configure interactively
upg config provider

Default Provider

# Set default provider
upg config set-default openai

# Switch to using Anthropic by default
upg config set-default anthropic

View Configuration

# Show current configuration
upg config show

Example output:

Current Configuration:
----------------------------------------
Default Provider: openai

Configured Providers:

OPENAI:
  Model: gpt-4o
  Temperature: 1.0
  API Key: sk-abcd...wxyz

ANTHROPIC:
  Model: claude-3-sonnet-20240229
  Temperature: 1.0
  API Key: sk-ant...4321

Configuration Storage

The tool stores configuration in ~/.config/upg/config.json:

• API keys for LLM providers
• Default provider settings
• Provider-specific configurations
• Saved prompts and their metadata

🗃️ Prompt Storage

Prompts are stored with:

• Unique name
• Description
• Variables list (if any)
• Tags for categorization
• Creation and update timestamps

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

⭐ Found this useful? Star the repo and share it!

Join Telegram Community