Generator Documentation - PromptAsCode

Overview

The Prompt Generator uses AI to help you create and improve prompts. Start from scratch with templates for common use cases, or improve existing prompts with intelligent suggestions. The generator understands prompt engineering best practices and applies them automatically.

AI-Powered

Intelligent generation

6+ Templates

Common use cases covered

Improve Mode

Enhance existing prompts

How to Use

1
Choose Your Mode - Select 'Create New' to start fresh or 'Improve Existing' to enhance a prompt you have.
2
Describe Your Goal - Enter what you want the prompt to accomplish. Be specific about the use case.
3
Select a Template (Optional) - For Create New mode, optionally choose a template that matches your use case.
4
Generate - Click Generate and the AI will create or improve your prompt based on best practices.
5
Refine & Iterate - Review the output, make adjustments, and regenerate if needed.

Generation Modes

Create New Mode

Generate a complete prompt from scratch. Describe what you need, and the AI will create a well-structured prompt following best practices.

Example Input

"I need a prompt for a customer support chatbot that handles refund requests. It should be empathetic, follow company policy, and know when to escalate."

Improve Existing Mode

Enhance a prompt you've already written. The AI analyzes your prompt and suggests improvements for clarity, structure, and effectiveness.

Improvements Include

Adding missing structure (role, task, format)
Clarifying ambiguous instructions
Adding constraints and guardrails
Optimizing for token efficiency
Adding examples where helpful

Templates

Templates provide proven starting points for common use cases:

Sales Assistant

Product recommendations, objection handling, lead qualification. Includes persuasion techniques and CRM integration points.

Code Review

Security analysis, best practices, performance suggestions. Structured output for integration with dev tools.

Blog Writer

SEO-optimized content, tone consistency, structure guidelines. Configurable for different content types.

Customer Support

Issue resolution, escalation protocols, empathy guidelines. Handles common scenarios with appropriate responses.

Data Analysis

Pattern recognition, insight generation, visualization suggestions. Works with various data formats.

Creative Writing

Story generation, character development, style consistency. Configurable genre and tone settings.

Customization

Generation Options

Verbosity: Choose concise, balanced, or detailed output
Structure: Include or exclude specific sections
Examples: Add few-shot examples automatically
Output Format: Specify expected response format
Constraints: Add safety guardrails and limitations

Advanced Options

For power users:

Specify target token count
Target specific AI models
Include domain-specific terminology
Add custom constraints and requirements

AI Expert Use Cases

Rapid Prototyping

Use the Generator to quickly create first-draft prompts for new features. Start with a template, customize, then refine through testing.

Best Practice Enforcement

Run existing prompts through Improve mode to ensure they follow best practices. Great for auditing legacy prompts or onboarding new content.

Team Training

Use generated prompts as examples for training team members on prompt engineering. The structured output demonstrates good practices.

Consistency at Scale

Generate prompts for similar use cases to maintain consistency across your application. Templates ensure all customer-facing prompts have the same quality.

Tips & Best Practices

Pro Tips

Be specific about your use case when describing goals
Include context about your target audience
Mention any constraints (token limits, format requirements)
Generate multiple variants and compare them
Always test generated prompts in the Playground
Use Improve mode on prompts that score low in the Linter

Iteration Workflow

Generate initial prompt
Test in Playground
Run through Linter
Use Improve mode for refinement
Repeat until satisfied
Save to Version control