Show HN: Open-source AI prompt engineering workbench with systematic evaluation

3 hours ago 1

The AI prompt engineering workbench that generates, analyzes, and systematically tests prompts

Stop writing prompts from scratch. Let AI help you craft better prompts, then systematically validate them. Built with Go for speed and reliability.

🎮 Try Interactive Demo

PromptForge Screenshot

Most prompt tools are glorified text editors. PromptForge brings engineering discipline to prompt development:

  • ✨ Smart Prompt Generation - AI-powered prompt crafting with intelligent suggestions
  • 🔍 Advanced Prompt Analysis - Get optimization feedback before you even test
  • 🧪 Systematic Evaluation - Generate comprehensive test suites automatically
  • 📊 Performance Analytics - Track what works across different scenarios
  • 📚 Version Control - Never lose a working prompt again

One-line Docker setup:

docker run -d -p 8080:8080 -e ANTHROPIC_API_KEY="your-key" promptforge:latest

Or clone and run:

git clone https://github.com/insaanimanav/promptforge.git cd promptforge && ./start.sh

Open http://localhost:8080 and start crafting better prompts.

✨ AI-Powered Prompt Generation

  • Smart Crafting: AI assists in building effective prompts from scratch
  • Intelligent Suggestions: Context-aware recommendations for prompt improvement

🔍 Advanced Prompt Analysis

  • Quick Review: Instant optimization suggestions
  • Deep Analysis: Comprehensive prompt evaluation and scoring
  • Best Practice Validation: Checks against proven prompt engineering principles

🧪 Systematic Evaluation Engine

Creates comprehensive test suites including:

  • Robustness: Edge cases, typos, input variations
  • Safety: Bias detection, harmful content resistance
  • Accuracy: Factual correctness verification
  • Creativity: Novel thinking scenarios
  • Execute with full parameter control
  • Dynamic variable detection and substitution
  • Multi-model comparison side by side (Claude, GPT-4, Azure OpenAI)
  • Organized library with search and tags
  • Complete execution history with filtering
  • Export/import capabilities
  • Claude 3.5 Sonnet (200K context) - Excellent reasoning
  • GPT-4.1 (200K context) - Detailed analysis
  • O3 (1M context) - Fast execution
  • Azure OpenAI - Enterprise-ready
# Anthropic (Default) -e ANTHROPIC_API_KEY="sk-ant-api03-..." # OpenAI -e OPENAI_API_KEY="sk-..." # Azure OpenAI -e AZURE_OPENAI_API_KEY="your-key" -e AZURE_OPENAI_BASE_URL="https://your-resource.openai.azure.com"
# Clone and setup git clone https://github.com/insaanimanav/promptforge.git cd promptforge # Configure environment cp .env.example .env # Add your API keys to .env # Run cd api && go run main.go

Core endpoints for integration:

  • POST /api/critique - Analyze prompts
  • POST /api/execute - Test with parameters
  • POST /api/generate-eval - Create test suites
  • GET /api/prompts - Manage prompt library

Built for the prompt engineering community. Contributions welcome:

  • 🌐 Additional AI model integrations
  • 📊 Enhanced evaluation metrics
  • 🎨 UI/UX improvements
  • 🔧 Performance optimizations
  1. Fork the repository
  2. Create feature branch: git checkout -b feature/name
  3. Submit pull request

GPLv3 - Built with ❤️ for better AI interactions


Transform prompt engineering from art to science

Read Entire Article