Creating a Prompt

This guide walks you through creating your first Prompt RDA.

What is a Prompt RDA?

A Prompt RDA uses large language models (LLMs) to generate text based on user inputs and a prompt template.

Prerequisites

xpay✦ Hub account
Understanding of prompt engineering
Ideas for valuable AI capabilities

Step 1: Define Your RDA

Basic Information

Name - Clear, descriptive title (e.g., “Competitor Analysis Report”)
Slug - URL-friendly identifier (e.g., “competitor-analysis-report”)
Description - What the RDA does (1-2 sentences)
Long Description - Detailed explanation (optional)

Step 2: Design Your Input Schema

Define what users need to provide:


const inputSchema = [
  {
    name: 'companyName',
    label: 'Company Name',
    type: 'text',
    required: true,
    placeholder: 'Acme Corp',
    description: 'The company to analyze'
  },
  {
    name: 'analysisDepth',
    label: 'Analysis Depth',
    type: 'select',
    required: true,
    options: ['basic', 'detailed', 'comprehensive']
  },
  {
    name: 'focusAreas',
    label: 'Focus Areas',
    type: 'text',
    required: false,
    placeholder: 'pricing, features, positioning'
  }
]

Field Types

Type	Description	Use For
`text`	Single-line input	Short text, names
`textarea`	Multi-line input	Long text, descriptions
`number`	Numeric input	Quantities, limits
`select`	Dropdown	Fixed options
`checkbox`	Boolean toggle	Flags, options
`url`	URL input	Links, endpoints
`file`	File upload	Documents

Step 3: Write Your Prompt Template

Create the prompt that will be sent to the LLM:


You are an expert business analyst.

Analyze the company {{companyName}} with {{analysisDepth}} depth.

{{#if focusAreas}}
Focus particularly on: {{focusAreas}}
{{/if}}

Provide:
1. Executive summary
2. Product/service overview
3. Competitive positioning
4. Strengths and weaknesses
5. Actionable recommendations

Format your response in clear markdown with headers and bullet points.

Template Syntax

{{variable}} - Insert variable value
{{#if var}}...{{/if}} - Conditional sections
{{#each items}}...{{/each}} - Loop over arrays

System Prompt (Optional)

Add a system prompt for consistent behavior:


You are a business analyst with deep experience in market research, competitive analysis, and strategic consulting.

Step 4: Configure Models

Select which models users can choose from:


const allowedModels = [
  {
    id: 'gpt-4o-mini',
    name: 'GPT-4o Mini',
    tier: 'fast',
    pricePerRun: 0.02
  },
  {
    id: 'claude-3.5-sonnet',
    name: 'Claude 3.5 Sonnet',
    tier: 'balanced',
    pricePerRun: 0.05
  },
  {
    id: 'o1-preview',
    name: 'o1 Preview',
    tier: 'reasoning',
    pricePerRun: 0.10
  }
]

Available Models

Fast Tier ($0.02)

GPT-4o Mini
Claude 3 Haiku
Llama 3.1 8B
Gemini Flash

Balanced Tier ($0.05)

GPT-4o
Claude 3.5 Sonnet
Llama 3.1 70B
Gemini Pro

Reasoning Tier ($0.10)

o1-preview
Claude 3 Opus
Llama 3.1 405B

Step 5: Add Examples

Create pre-generated examples to showcase your RDA:


const examples = [
  {
    id: 'ex_001',
    title: 'SaaS Competitor Analysis',
    inputs: {
      companyName: 'Salesforce',
      analysisDepth: 'detailed',
      focusAreas: 'pricing, enterprise features'
    },
    output: '## Competitor Analysis\n\n...',
    model: 'claude-3.5-sonnet'
  }
]

Step 6: Set Pricing

Per-Run Price

Included in model tier pricing above.

Source Code Price (Optional)

Allow users to buy your prompt template:


const promptSourcePrice = 4.99 // USDC

Step 7: Submit for Review

Before publishing:

Test thoroughly - Run with various inputs
Check output quality - Ensure consistent results
Review pricing - Compare to similar RDAs
Add examples - Show best use cases

Submit your RDA for verification review.

Best Practices

Prompt Quality

Be specific about output format
Include examples in the prompt
Handle edge cases gracefully

User Experience

Use clear field labels
Provide helpful descriptions
Show placeholder examples

Pricing

Start competitive
Consider your costs
Offer value vs alternatives