Prompt Piper CLI

A command-line tool for compressing AI prompts to reduce token usage and save costs.

Two Modes

1. Command Mode (Quick Operations)

Run one-off compressions directly from the command line.

2. Interactive Mode (Terminal UI)

Full terminal interface with menu-driven options, similar to Claude.

Installation

From NPM (coming soon)

npm install -g prompt-piper-cli

Local Development

cd cli
npm install
npm run build
npm link  # Makes 'prompt-piper' available globally

Quick Start

Interactive Mode (Recommended for demos)

# Start interactive terminal UI
prompt-piper interactive

# Or use the shortcut
ppi

Command Mode

# Compress a prompt directly
prompt-piper compress "Could you please help me write detailed code?"

# From file
prompt-piper compress -f my-prompt.txt -o compressed.txt

# Pipe input
echo "verbose prompt here" | prompt-piper compress -q

# Analyze without compressing
prompt-piper analyze "My long prompt here"

# Side-by-side comparison
prompt-piper compare -f verbose.txt

# Show examples
prompt-piper examples

Interactive Mode

The interactive mode provides a full terminal UI experience:

╔═══════════════════════════════════════════════════════════════════════════════╗
║   ██████╗ ██████╗  ██████╗ ███╗   ███╗██████╗ ████████╗                      ║
║   ██╔══██╗██╔══██╗██╔═══██╗████╗ ████║██╔══██╗╚══██╔══╝                      ║
║   ██████╔╝██████╔╝██║   ██║██╔████╔██║██████╔╝   ██║                         ║
║   ██╔═══╝ ██╔══██╗██║   ██║██║╚██╔╝██║██╔═══╝    ██║                         ║
║   ██║     ██║  ██║╚██████╔╝██║ ╚═╝ ██║██║        ██║                         ║
║   ╚═╝     ╚═╝  ╚═╝ ╚═════╝ ╚═╝     ╚═╝╚═╝        ╚═╝                         ║
╚═══════════════════════════════════════════════════════════════════════════════╝

MAIN MENU
────────────────────────────────────────
1) Compress Example 1: Verbose API Request
2) Compress Example 2: Over-polite Coding Request
3) Compress Example 3: Redundant Instructions

4) Enter Custom Prompt
5) View Compression Stats

H) Help
C) Clear Screen
Q) Quit
────────────────────────────────────────

prompt-piper>

Features:

• Menu-driven interface - Simple numbered options
• Built-in examples - Test compression instantly
• Custom prompt entry - Enter multi-line prompts
• Session statistics - Track total savings
• Visual compression - See before/after with token bars
• Help system - Built-in documentation

Usage:

# Start interactive mode
prompt-piper interactive

# Or use shortcut
ppi

# Navigate with:
# - Numbers (1-5) for options
# - H for help
# - C to clear screen
# - Q to quit

Commands (Command Mode)

compress (alias: c)

Compress a prompt to reduce token usage.

Options:

• -f, --file <path> - Read prompt from file
• -o, --output <path> - Save compressed prompt to file
• -q, --quiet - Only output the compressed prompt
• -s, --stats - Show detailed compression statistics

Examples:

prompt-piper compress "Please help me code"
prompt-piper c -f prompt.txt -s
prompt-piper compress -q "verbose text" > output.txt

analyze (alias: a)

Analyze a prompt to see potential compression savings.

Options:

• -f, --file <path> - Read prompt from file

Example:

prompt-piper analyze -f my-long-prompt.txt

examples (alias: ex)

Show example usage and sample compression results.

prompt-piper examples

Usage Patterns

1. Interactive Development

# Quick compression
prompt-piper compress "Could you please help me write a very detailed guide?"

# Output: Write detailed guide.
# Stats: 68 → 19 tokens (-49 | 72.1%)

2. File Processing

# Process files
prompt-piper compress -f prompts/verbose.txt -o prompts/compressed.txt -s

# Batch processing
for file in prompts/*.txt; do
  prompt-piper compress -f "$file" -o "compressed/$(basename "$file")"
done

3. Pipeline Integration

# Part of a larger workflow
curl -s api.example.com/prompt | prompt-piper compress -q | ai-tool --stdin

4. Cost Analysis

# Analyze potential savings
prompt-piper analyze -f expensive-prompt.txt

# Output shows:
# Current tokens: 150
# After compression: 89 tokens (-61 tokens, 40.7% reduction)
# Cost savings: $0.0018 per request

Sample Output

Standard Compression

🎵 Prompt Piper Results
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📝 Compressed Prompt:
Write stepwise guide implementing React authentication component.

Stats: 147 → 87 tokens (-60 | 40.8%)

Detailed Stats (-s flag)

📊 Compression Stats:
Original tokens: 147
Compressed tokens: 87
Tokens saved: 60
Compression ratio: 40.8%
Cost savings: $0.0018 per request

Analysis Mode

🎵 Prompt Analysis
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📊 Current Stats:
Current tokens: 147
Estimated cost: $0.0044 per request

💡 Potential Savings:
After compression: 87 tokens
Tokens saved: 60
Compression ratio: 40.8%
Cost savings: $0.0018 per request

💰 Run prompt-piper compress to apply compression!

How It Works

The CLI tool uses intelligent pattern matching to:

• Remove redundant phrases (please, could you, I would like)
• Compress verbose instructions (step by step → stepwise)
• Eliminate filler words (very, really, quite)
• Clean up excessive whitespace
• Preserve meaning and intent

Integration Examples

Shell Scripts

#!/bin/bash
# compress-and-send.sh
COMPRESSED=$(prompt-piper compress -q -f "$1")
echo "$COMPRESSED" | curl -X POST api.openai.com/chat -d @-

CI/CD Pipeline

- name: Optimize Prompts
  run: |
    find prompts/ -name "*.txt" -exec \
      prompt-piper compress -f {} -o optimized/{} \;

Development Workflow

# Add to your shell profile
alias compress-prompt='prompt-piper compress -s'
alias analyze-prompt='prompt-piper analyze'

Tips for Best Results

1. Use with longer prompts - Short prompts may not compress much
2. Check with -s flag - Verify compression maintains meaning
3. Test compressed prompts - Ensure AI response quality
4. Batch process - Use file options for multiple prompts
5. Combine with cost monitoring - Track actual API savings

Made for hackathons, optimized for production! 🚀