Prompter Language

Prompter Language is a domain-specific language (DSL) designed to bring structure, type safety, and reusability to AI prompt engineering. It provides a declarative way to define, validate, and execute prompts for AI language models.

Overview

Think of Prompter Language as "TypeScript for prompts" - it adds static typing, validation, and structured composition to what would otherwise be plain text prompts. Key aspects include:

Declarative Syntax

prompt NewsArticle {
  source: url("https://api.news.com/data") {
    depth: 1
  }
  variables {
    title: string;
    tone: "neutral" | "critical";
  }
  validation {
    title: required, minLength(10);
  }
}

Type Safety

• Strong typing for variables and parameters
• Union types for constrained choices
• Runtime validation of inputs
• Compile-time checking of prompt structure

Content Pipeline

1. Source: Fetch content from URLs, files, or Git repos
2. Variables: Define and validate input parameters
3. Processing: Transform content before/after AI processing
4. Output: Format and structure the final result

Integration Features

• Built-in support for OpenAI and other AI providers
• Extensible source system for content fetching
• Markdown and other output formats
• Version control friendly

The language aims to solve common prompt engineering challenges:

• Inconsistent prompt formatting
• Lack of input validation
• Difficulty in maintaining prompt versions
• Poor reusability across projects
• Missing type safety
• Ad-hoc content processing

Features

• 🎯 Semantic Prompt Definition: Define prompts using a clear, structured language
• 🔍 Source Integration: Support for multiple content sources (URL, File, Git)
• ✅ Variable Validation: Type checking and validation for prompt variables
• 📝 Multiple Output Formats: Support for Markdown and other output formats
• 🔄 Processing Pipeline: Pre and post-processing hooks for content
• 🎨 Style Control: Fine-grained control over tone, style, and formatting

Installation

# Install using npm
npm install prompter-language

# Or using yarn
yarn add prompter-language

Quick Start

1. Create a .env file with your API keys:

OPENAI_API_KEY=your_openai_api_key_here

2. Create a prompt file (e.g., article.prompt):

prompt NewsArticle {
  source: url("https://example.com/news") {
    depth: 0
  }

  variables {
    title: string;
    summary: string;
    tone: "neutral" | "analytical" | "critical";
    style: "news" | "editorial" | "feature";
  }

  validation {
    title: required;
    summary: required;
    tone: required;
    style: required;
  }

  output {
    format: "markdown"
  }
}

3. Run the prompt:

yarn start path/to/your/prompt.file

Prompt Language Syntax

Source Types

• URL Source:

  source: url("https://example.com") {
    depth: 0,
    timeout: 5000
  }

• File Source:

  source: file("./template.txt") {
    format: "text",
    encoding: "utf-8"
  }

• Git Source:

  source: git("https://github.com/user/repo.git") {
    branch: "main",
    depth: 1
  }

Variables

variables {
  name: string;
  age: number;
  tags: string[];
  metadata: {
    created: string;
    author: string;
  };
}

Validation

validation {
  name: required, minLength(2), maxLength(50);
  age: required, min(0), max(150);
  tags: required, minItems(1);
}

Output Configuration

output {
  format: "markdown";
  maxLength: 2000;
  temperature: 0.7;
}

Development

Prerequisites

• Node.js (v16 or higher)
• Yarn or npm
• OpenAI API key

Setup

1. Clone the repository:

git clone https://github.com/yourusername/prompter-language.git
cd prompter-language

2. Install dependencies:

yarn install

3. Build the project:

yarn build

4. Run tests:

yarn test

Project Structure

src/
  ├── cli/           # Command-line interface
  ├── parser/        # Prompt language parser
  ├── services/      # AI service integrations
  ├── sources/       # Content source handlers
  ├── types/         # TypeScript type definitions
  └── examples/      # Example prompts

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'feat: add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

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

Acknowledgments

• OpenAI for their GPT models and API
• PeggyJS for the parser generator
• All contributors who have helped shape this project