A Val Town application that formats X.com (formerly Twitter) article-style posts in a clean, aesthetically pleasing reading experience. The application fetches content from the FXTwitter API and displays it with minimal black and white design, optimized for readability.

• Pattern: x.com/user/status/id → subdomain.val.run/user/status/id
• Support: Accept both x.com and twitter.com domains
• Normalization: Internally normalize all URLs to x.com format before processing

• Primary Interface: Interactive web page with URL input form
• Homepage: Landing page with description, input form, and instructional examples showing the URL pattern
• Example Content: Balance between showing example threads and clear instructions on how to use the service

• Target Content: Article threads only (long-form thread articles)
• Article Format: API provides article content in a single structure (not individual threaded tweets)
• Display: Single continuous article view

• Style: Minimal black and white design
• Purpose: Distraction-free, focus entirely on readability and content
• Color Palette: Pure black text on white background, no dynamic theming from API color data

• Font Family: Modern sans-serif (system fonts for fast loading)
• Font Stack: system-ui, -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif
• Content Width: Maximum ~650px for optimal readability (65-75 characters per line)
• Responsive: Centered content with padding on larger screens

1. Hero Image: Cover media displayed as large hero image at top
2. Article Title: Large H1 heading below hero image
3. Author Information: Display name + avatar image
4. Metadata: Publication date (relative format: "2 hours ago", "yesterday" for recent; full date for older)
5. Article Content: Full formatted content with proper spacing
6. Footer: Link to view original on X.com

• Rendering Style: Full formatting support
• Supported Elements:
  • Headers (H2/H3 for header-two, etc.)
  • Unordered and ordered lists (<ul>, <ol>)
  • Bold and italic text via inline style ranges
  • Code blocks with syntax highlighting
  • Links (converted from entity ranges)
  • Paragraphs with proper spacing

• Support Level: Full entity support
• Entity Types:
  • URLs → clickable links with appropriate attributes
  • Code blocks → properly formatted with syntax highlighting
  • Mentions → formatted appropriately
  • All entity ranges from entityMap processed correctly

• Images: Full media display inline within article
• Cover Image: Displayed as prominent hero image
• Embedded Media: All images from content shown at full width within content container
• Loading: Load all media immediately (no lazy loading)

• Library: Highlight.js
• Approach: Load core library + only needed languages for minimal bundle size
• Theme: Minimal, monochrome theme matching the black/white aesthetic
• Features: Basic syntax highlighting without line numbers or extra UI

• Type: HTTP Val (fileType: "http")
• Runtime: Deno serverless environment
• Framework: Hono for routing and API endpoints

GET /                           → Homepage with input form
GET /:username/status/:id       → Formatted article view

1. Parse URL to extract username and status ID
2. Check SQLite cache for existing article data
3. If not cached or expired (>24 hours), fetch from FXTwitter API
4. Store response in SQLite cache with timestamp
5. Parse API response and render HTML
6. Return formatted article page

• Backend: Hono for routing
• Frontend: Highlight.js for syntax highlighting (minimal build)
• API: FXTwitter API (https://api.fxtwitter.com/:username/status/:id)
• Styling: Pure CSS, no frameworks
• JavaScript: None on article pages (pure static HTML)

CREATE TABLE IF NOT EXISTS articles (
  tweet_id TEXT PRIMARY KEY,
  fetched_at INTEGER NOT NULL,
  data TEXT NOT NULL
);

• Key: Tweet status ID (unique identifier)
• TTL: 24 hours
• Expiration Logic: Refetch articles older than 24 hours
• Storage Format: JSON string of complete FXTwitter API response
• Invalidation: Time-based only (no manual invalidation)

// Check cache
const cached = await sqlite.execute({
  sql: "SELECT data, fetched_at FROM articles WHERE tweet_id = ?",
  args: [tweetId]
});

// Is cache valid?
const cacheAge = Date.now() - cached.rows[0].fetched_at;
const isExpired = cacheAge > 24 * 60 * 60 * 1000; // 24 hours

// Store in cache
await sqlite.execute({
  sql: "INSERT OR REPLACE INTO articles (tweet_id, fetched_at, data) VALUES (?, ?, ?)",
  args: [tweetId, Date.now(), JSON.stringify(apiResponse)]
});

• Endpoint: https://api.fxtwitter.com/:username/status/:id
• Response Structure: See example at https://api.fxtwitter.com/francedot/status/2015178880215298557

interface APIResponse {
  code: number;
  message: string;
  tweet: {
    url: string;
    id: string;
    author: {
      name: string;
      screen_name: string;
      avatar_url: string;
    };
    article: {
      title: string;
      created_at: string;
      cover_media: {
        url: string;
        width: number;
        height: number;
      };
      content: {
        blocks: Array<{
          type: string;  // "unstyled", "header-two", "unordered-list-item", etc.
          text: string;
          inlineStyleRanges: Array<{
            offset: number;
            length: number;
            style: string;  // "BOLD", "ITALIC", etc.
          }>;
          entityRanges: Array<{
            offset: number;
            length: number;
            key: number;
          }>;
        }>;
        entityMap: {
          [key: string]: {
            type: string;  // "LINK", "CODE", etc.
            data: any;
          };
        };
      };
    };
  };
}

1. Invalid URL: Not a valid X.com/Twitter.com article URL
2. Deleted Tweet: FXTwitter returns 404
3. Private Account: FXTwitter returns access error
4. API Failure: FXTwitter API is down or unresponsive
5. Not an Article: Tweet exists but is not an article-style post

• UI: Clean, minimal error page
• Content:
  • Clear explanation of what went wrong
  • Link to go back to original article on x.com
  • Consistent styling with main design
• HTTP Status: Appropriate status codes (404, 500, etc.)
• No Retry Logic: Simple failure, no automatic retries
• Stale Cache: Do not serve expired cached content on API failure

<div class="error-container">
  <h1>Unable to load article</h1>
  <p>[Contextual error message]</p>
  <a href="[original x.com URL]">View on X.com</a>
</div>

1. Service Title/Logo: Clear branding for the service
2. Description: Brief explanation of what the service does
3. URL Pattern Example: Show the pattern: x.com/user/status/id → subdomain.val.run/user/status/id
4. Input Form: URL input field with submit button
5. Example Links: 1-2 example article URLs demonstrating the service

• Input: Text field accepting X.com or Twitter.com URLs
• Validation: Check if URL matches expected pattern
• Action: Submit button that redirects to formatted article path
• Transformation: Extract username and status ID, navigate to /:username/status/:id

User pastes: https://x.com/francedot/status/2015178880215298557
Form extracts: username=francedot, id=2015178880215298557
Redirects to: /francedot/status/2015178880215298557

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  <title>[Article Title] - [Author Name]</title>

  <!-- OpenGraph tags -->
  <meta property="og:title" content="[Article Title]">
  <meta property="og:description" content="[Preview text or first paragraph]">
  <meta property="og:image" content="[Cover media URL]">
  <meta property="og:type" content="article">

  <!-- Styles -->
  <link rel="stylesheet" href="[styles.css or inline]">

  <!-- Highlight.js for syntax highlighting -->
  <link rel="stylesheet" href="[highlight.js theme]">
  <script src="[highlight.js core + languages]"></script>
</head>
<body>
  <article class="article-container">
    <img src="[cover_media.url]" alt="Cover image" class="hero-image">

    <header class="article-header">
      <h1 class="article-title">[Article Title]</h1>

      <div class="author-info">
        <img src="[author.avatar_url]" alt="[author.name]" class="avatar">
        <span class="author-name">[author.name]</span>
      </div>

      <time class="publish-date">[Relative date]</time>
    </header>

    <div class="article-content">
      [Rendered blocks with full formatting]
    </div>

    <footer class="article-footer">
      <a href="[original tweet URL]" target="_blank" rel="noopener">
        View original on X.com
      </a>
    </footer>
  </article>
</body>
</html>

const blockTypeMap = {
  'unstyled': 'p',
  'header-one': 'h1',
  'header-two': 'h2',
  'header-three': 'h3',
  'unordered-list-item': 'li',  // wrap in <ul>
  'ordered-list-item': 'li',     // wrap in <ol>
  'atomic': 'div',               // special handling for embedded content
  'code-block': 'pre'            // wrap in <code> with highlighting
};

function applyInlineStyles(text: string, ranges: InlineStyleRange[]): string {
  // Sort ranges by offset
  // Apply BOLD → <strong>, ITALIC → <em>, etc.
  // Handle overlapping ranges correctly
  // Return HTML string with inline formatting
}

function applyEntityRanges(text: string, ranges: EntityRange[], entityMap: EntityMap): string {
  // For each entity range:
  // - Extract entity from entityMap by key
  // - Apply appropriate transformation:
  //   - LINK → <a href="...">text</a>
  //   - CODE → <code>text</code>
  //   - Handle other entity types as needed
  // Return HTML string with entities processed
}

• Load Everything: Render complete article at once (no pagination or lazy loading)
• Minimal Dependencies: Only essential libraries loaded
• SQLite Cache: Reduces API calls and improves response time for repeated views
• Static HTML: No client-side JavaScript on article pages (except Highlight.js initialization)
• Content Width: Fixed max-width prevents excessive line length and maintains readability

• Core HTML/CSS: < 10KB
• Highlight.js: < 50KB (core + common languages only)
• Total Initial Load: < 100KB excluding images
• Time to First Byte: < 200ms (cached), < 1s (uncached)

• No Tracking: No analytics or user behavior logging
• No Cookies: No cookies set or required
• No Third-Party Scripts: Only self-hosted or CDN resources (Highlight.js)
• No User Data Storage: Only article content cached, no user identifiers

• Application Logs: Server-side errors logged for debugging
• No Access Logs: Do not log article views or user behavior
• Error Tracking: Only log failed API requests for debugging purposes

<meta property="og:title" content="[Article Title]">
<meta property="og:description" content="[Article preview_text or excerpt]">
<meta property="og:image" content="[cover_media.url]">
<meta property="og:type" content="article">
<meta property="og:url" content="[Current URL]">

<title>[Article Title] by [Author Name]</title>

function formatDate(timestamp: string): string {
  const date = new Date(timestamp);
  const now = new Date();
  const diffInHours = (now - date) / (1000 * 60 * 60);

  if (diffInHours < 1) return "Just now";
  if (diffInHours < 24) return `${Math.floor(diffInHours)} hours ago`;
  if (diffInHours < 48) return "Yesterday";
  if (diffInHours < 168) return `${Math.floor(diffInHours / 24)} days ago`;

  // For older articles, show full date
  return date.toLocaleDateString('en-US', {
    year: 'numeric',
    month: 'long',
    day: 'numeric'
  });
}

• Set up Hono app with route handlers
• Implement SQLite schema and caching logic
• Create FXTwitter API client with error handling
• Build URL parser for both x.com and twitter.com
• Implement cache lookup and storage
• Create article renderer (blocks → HTML)
• Build error page template

• Design minimal CSS stylesheet
• Create homepage HTML template
• Build article page HTML template
• Integrate Highlight.js for syntax highlighting
• Implement inline style ranges rendering
• Implement entity ranges rendering
• Add OpenGraph meta tags
• Style error pages

• Test with various article types
• Verify cache expiration logic
• Test error scenarios (404, API down, etc.)
• Validate URL parsing for both domains
• Check rendering of all block types
• Verify syntax highlighting works
• Test social sharing metadata
• Mobile responsiveness testing

• Deploy to Val Town
• Verify endpoint URL
• Test live with real X.com articles
• Monitor for errors
• Document usage

• Dark mode toggle
• Reader preferences (font size, width)
• Article bookmarking/favorites
• RSS feed generation
• Print-optimized styling
• Reading time estimates
• Multi-language support
• Search functionality
• Article collections/categorization

1. Functionality: Successfully formats X.com article threads with proper rendering
2. Performance: Pages load in < 2 seconds including images
3. Readability: Clean, distraction-free reading experience
4. Reliability: Graceful error handling for all edge cases
5. Simplicity: Minimal dependencies, straightforward codebase
6. Privacy: No tracking or unnecessary data collection