AI Everywhere

ONE makes it easy to add AI into your app using markdown, yaml, and typescript. Add AI to any type of content - pages, docs, blog posts, lessons, videos. You can use eternal data from anywhere like Shopify, Wordpress or Notion.

Hierarchy

The system follows a cascading inheritance pattern where configurations and prompts flow from base to specific implementations:

1. Core System Level (src/1/1.md)

• Here you will find the core system prompt to build the system
• Establishes fundamental AI behavior patterns
• Contains base feature explanations and agent characteristics

2. I (src/1/I.md)

• Appends business-specific instructions and knowledge
• Customizes AI behavior for specific use cases
• Allows for domain-specific rules and responses

3. Chat src/components/Chat.tsx Assembles prompt s

3. Layout Level (src/layouts/Layout.astro)

• Provides default chat configurations that fall ba
• Sets up standard UI components and behaviors
• Establishes baseline prompt settings

4. Page Level Configuration

• page.astro: Override layout settings when needed
• page.md: Custom content-specific configurations
• Docs.astro: Documentation-specific settings

Configuration Flow

graph TD
    A[Base System - 1.md] --> B[Business Custom - I.md]
    B --> C[Layout.astro]
    C --> D[Page Level]
    D --> E[page.astro]
    D --> F[page.md]
    D --> G[Docs.astro]

Implementation Details

1. Base Configuration

// Default chat configuration in Layout.astro
const defaultChatConfig = ChatConfigSchema.parse({
  provider: 'mistral',
  model: 'mistral-large-latest',
  systemPrompt: [{
    type: 'text',
    text: 'I am Agent ONE. How can I help you today?'
  }]
});

2. Page-Level Override

---
const chatConfig = ChatConfigSchema.parse({
  systemPrompt: [{
    type: "text",
    text: "Custom prompt for this page"
  }],
  welcome: {
    message: "Welcome to this section!",
    suggestions: []
  }
});
---

<Layout chatConfig={chatConfig}>
  <!-- Page content -->
</Layout>

Best Practices

1. Modular Configuration

   • Keep base prompts focused and reusable
   • Use business customization for specific domains
   • Override only necessary settings at each level

2. Inheritance Management

   • Follow the configuration cascade
   • Document overrides clearly
   • Maintain consistent behavior patterns

3. Performance Optimization

   • Configure appropriate token limits
   • Set temperature based on use case
   • Choose suitable models for different scenarios

Advanced Usage

Custom Endpoints

const config = {
  provider: 'custom',
  apiEndpoint: 'https://your-api.com',
  apiKey: process.env.API_KEY
};

Runtime Selection

const config = {
  runtime: 'edge', // or 'node'
  model: 'mistral-large-latest'
};

Next Steps

1. Review the base system prompt in src/1/1.md
2. Customize business rules in src/1/I.md
3. Configure layout defaults in Layout.astro
4. Implement page-specific overrides as needed