prompt-enhancer

CLI/TUI tool to enhance prompts using various LLM providers (Gemini, Copilot, Kilo-gateway, Opencode Zen)

Status: ✅ MVP Complete - Phase 4 Interactive TUI Ready for Use

Features

• 🎯 Interactive TUI Mode - Menu-driven interface for prompt enhancement
• ⚡ CLI Mode - Direct enhancement via command-line flags
• 🔧 Multi-Provider Support - Gemini, Copilot, Kilo-gateway
• 💾 Config Management - Persistent settings at ~/.prompt-enhancer/
• 📜 History Tracking - Browse past enhancements
• 🔄 Streaming Output - Real-time enhanced prompt display
• 📋 Post-Enhancement Actions - Copy to clipboard, enhance further, or return to main menu
• 🧠 Enhance Further Presets - Make concise, make more detailed, add structure/output format, improve for coding tasks
• 🎨 Interactive Settings - Configure providers via TUI

Install

npm install --global prompt-enhancer

Or use directly:

npx prompt-enhancer

Quick Start

Interactive TUI Mode (Default)

prompt-enhancer

This launches the interactive menu where you can:

• Enhance a prompt
• Configure provider settings
• View enhancement history
• Get help

After an enhancement completes in TUI mode, you can:

• Copy the enhanced prompt to clipboard
• Run "Enhance further" with presets
• Return to the main menu without exiting

CLI Mode - Direct Enhancement

$ prompt-enhancer --prompt "Your prompt here"
# or shorter
$ prompt-enhancer -p "Your prompt here"

Configuration

Setting Up Providers

1. Via TUI Settings Menu (Interactive):

   $ prompt-enhancer
   # Select "Settings" → Choose provider → Enter API Key → Select Model

2. Manual Config File: Edit ~/.prompt-enhancer/config.json:

   {
   	"version": "0.1.0",
   	"providers": {
   		"gemini": {
   			"name": "gemini",
   			"apiKey": "your-api-key",
   			"model": "gemini-2.5-flash",
   			"enabled": true
   		},
   		"copilot": {
   			"name": "copilot",
   			"apiKey": "your-api-key",
   			"model": "gpt-4o",
   			"enabled": true
   		},
   		"kilo": {
   			"name": "kilo",
   			"apiKey": "your-api-key",
   			"model": "auto",
   			"enabled": true
   		}
   	},
   	"defaultProvider": "gemini",
   	"maxTokens": 500,
   	"temperature": 1.0,
   	"saveHistory": true,
   	"streaming": true,
   	"theme": "dark"
   }

Usage Examples

Interactive Enhancement

$ prompt-enhancer
# Select "Enhance Prompt" → Type your prompt → Choose provider → View result

Quick CLI Enhancement

prompt-enhancer -p "write a hello world program"

View Enhancement History

$ prompt-enhancer
# Select "View History" to browse past enhancements

Get Help

$ prompt-enhancer --help
$ prompt-enhancer
# Select "Help" in menu for interactive guide

Supported Models

Gemini

• gemini-2.5-flash (default)
• gemini-2.0-pro
• gemini-1.5-flash
• gemini-1.5-pro

Copilot / OpenAI

• gpt-4o (default)
• gpt-4.1
• gpt-4-turbo
• gpt-3.5-turbo

Kilo Gateway

• auto (default) - Automatically routes to best available free model
• Any model supported by Kilo's router

Configuration Files

• Config: ~/.prompt-enhancer/config.json - Provider settings, API keys, preferences
• History: ~/.prompt-enhancer/history.json - Enhancement history with timestamps

API Keys Setup

Gemini

Get your API key at: https://makersuite.google.com/app/apikey

OpenAI / Copilot

Get your API key at: https://platform.openai.com/api-keys

Kilo Gateway

Get your key at: https://kilo.dev (or use free models with no key)

Commands

$ prompt-enhancer --help

  Usage
    $ prompt-enhancer
    $ prompt-enhancer --prompt "your prompt"
    $ prompt-enhancer enhance "your prompt"

  Options
    --prompt, -p              Prompt text for headless mode
    --output, -o              Write output to a file
    --output-format           Output format: auto | txt | json | md (default: auto)
    --verbose                 Verbosity level: 1 | 2 | 3 (default: 1)
    --debug                   Enable debug diagnostics + API request/response logs
    --trace                   Enable step-by-step execution trace
    --about                   Show package metadata
    --version, -v             Show version
    --help, -h                Show help

  Subcommands
    help                      Show comprehensive help
    about                     Show package metadata
    version                   Show version details
    enhance                   Run headless enhancement using positional prompt text

  Examples
    $ prompt-enhancer
    $ prompt-enhancer -p "build a cli to enhance prompts"
    $ prompt-enhancer -p "summarize this architecture" -o .\out\result.json --output-format json
    $ prompt-enhancer enhance "design API retries" --debug --trace --verbose 3
    $ prompt-enhancer --help

Architecture

Project Structure

src/
├── cli.tsx                 # CLI entry point and routing
├── app.tsx                 # Interactive TUI main app
├── commands/               # Command implementations
│   ├── direct-enhance.tsx  # CLI enhancement
│   ├── about.tsx
│   ├── help.tsx
│   └── version.tsx
├── components/             # React/Ink TUI components
│   ├── enhance-prompt.tsx  # Interactive enhancement UI
│   ├── settings.tsx        # Provider configuration UI
│   ├── history-viewer.tsx  # History browsing UI
│   └── menu.tsx            # Menu component
└── lib/                    # Core libraries
    ├── providers/          # Provider implementations
    │   ├── gemini.ts
    │   ├── copilot.ts
    │   ├── kilo.ts
    │   ├── base.ts         # Abstract base provider
    │   └── index.ts        # Provider registry
    ├── config/             # Configuration management
    │   ├── manager.ts      # Config read/write
    │   ├── schema.ts       # Zod validation
    │   └── types.ts        # TypeScript types
    ├── enhancement/        # Enhancement engine
    │   └── engine.ts       # Streaming enhancement logic
    ├── history/            # History management
    │   └── manager.ts      # History persistence
    └── utils/              # Utilities

Development

Build

bun run build

Lint & Format

bun run lint
bun run format

Type Check

bun run typecheck

Full Build Pipeline

bun run build  # Runs format, lint, typecheck, then compile

Technology Stack

• React 18 - Component framework
• Ink 5.2.1 - Terminal UI rendering
• TypeScript - Type safety
• Zod - Schema validation
• Meow - CLI argument parsing
• Provider SDKs:
  • @google/generative-ai - Gemini API
  • openai - OpenAI/Copilot API
  • Custom HTTP client for Kilo

Notes

• Configuration is stored locally in ~/.prompt-enhancer/
• API keys are stored unencrypted in config.json (consider using environment variables for production)
• History is automatically saved unless disabled in config
• Streaming output requires provider and model support

License

MIT