Prompt Alchemy

Transform raw ideas into golden prompts through the ancient art of linguistic alchemy

Prompt Alchemy

Prompt Alchemy transforms your raw ideas into high-quality prompts for AI systems through a three-phase alchemical process, featuring both a powerful Go backend and a beautiful React frontend.

The tool uses a three-phase refinement process to create precise, effective prompts. You can use it as a command-line tool, integrate it with AI agents, or interact through the web UI.

What you’ll accomplish: Generate better prompts, integrate with multiple AI providers, and improve results over time with machine learning.

Key features

Three-phase prompt refinement: Improves your prompts through idea extraction, natural language flow, and precision tuning
Multiple AI providers: Works with OpenAI, Anthropic, Google, OpenRouter, and local Ollama models
Intelligent selection: Uses AI to automatically choose the best prompt variants
Flexible deployment: Run as a CLI tool, API server, or web application
Beautiful UI: React frontend with 3D visualizations and alchemy-themed design
Adaptive learning: Improves recommendations over time based on your feedback
Local storage: Keeps all your data in a local SQLite database for privacy and speed
Docker support: Full containerization with docker-compose profiles

Quick start

Generate a prompt (CLI)

# Generate a prompt with all three refinement phases
prompt-alchemy generate "A blog post about the future of AI" --persona=writing

Start the web UI

# Using Docker (recommended)
docker-compose --profile hybrid up -d

# Or for development with hot reload
npm install
npm run dev

Access the web UI at http://localhost:8080

Start a server (for AI agents)

# Start the MCP server for AI agent integration
prompt-alchemy serve

AI agents can connect to the server using stdin/stdout and make JSON-RPC calls.

How it works

Prompt Alchemy refines your ideas through three phases:

Extract core concepts: Identifies the key ideas from your input
Create natural flow: Transforms concepts into readable, flowing language
Add precision: Refines the prompt for accuracy and effectiveness

Each phase can use a different AI provider for optimal results.

Why use Prompt Alchemy?

Unlike manual prompt engineering, Prompt Alchemy provides a consistent, measurable approach:

Repeatable process: Every prompt goes through the same proven refinement steps
Provider flexibility: Use the best AI model for each phase of improvement
Performance tracking: Monitor success rates and identify what works best
Continuous improvement: The system learns from your feedback to get better over time

Documentation

Getting Started

Getting Started - Installation and first steps
Installation Guide - Detailed setup instructions
Usage Guide - Command reference and examples

🔄 Operational Modes

On-Demand vs Server Mode - Comprehensive comparison of operational modes
Mode Quick Reference - Quick decision guide and command reference
Mode Selection FAQ - Frequently asked questions about choosing a mode.
Deployment Guide - Complete deployment strategies for both modes.

Technical Documentation

Architecture - Technical design and internals
Database - Database schema and implementation details
Vector Embeddings - Semantic search and vector storage implementation
Diagrams - Visual architecture and flow diagrams

Server Mode & Integration

MCP Integration - Model Context Protocol server setup
MCP API Reference - Detailed reference for all 6 MCP tools.
Learning Mode - Adaptive learning configuration
HTTP API Reference - RESTful API endpoints and models

Development & Operations

CLI Reference - Complete command-line interface documentation
Automated Scheduling - Set up nightly training jobs with cron/launchd
Multi-Arch Builds - Cross-platform build system and CI/CD
Renovate Setup - Automated dependency updates
Release Automation - Semantic versioning and GitHub releases

Support

Issues: GitHub Issues
Contributing: See CONTRIBUTING.md

License

Prompt Alchemy is open source software licensed under the MIT License.