Third Audience

Serve AI-optimized Markdown to LLM crawlers. Track who's citing your content. Get discovered by ChatGPT, Perplexity, Claude, and Gemini.

---

What Is Third Audience?

For two decades, websites were built for two audiences: humans and search engines. Today there is a third audience rapidly growing: AI agents and LLM crawlers.

AI systems like ChatGPT, Perplexity, Claude, and Google Gemini actively crawl the web to answer questions, generate summaries, and provide recommendations. Most websites serve these bots the same HTML built for humans — full of navigation menus, scripts, ads, and layout markup that adds noise and reduces accuracy.

Third Audience automatically serves clean, structured Markdown versions of your content to AI crawlers, while humans continue to see your normal website. Think of it as SEO for the AI era — Generative Engine Optimization (GEO).

---

How It Works

AI Bot visits yoursite.com/page-name.md
        │
        ▼
Plugin intercepts the request
        │
        ▼
Bot Detection Pipeline
  ├── Known Pattern Detector (database lookup)
  └── Heuristic Detector (user agent rules)
        │
        ▼
HTML → Markdown conversion (local, no external APIs)
        │
        ▼
Cached Markdown served (1–4ms response time)
        │
        ▼
Visit recorded in analytics dashboard

1. AI crawlers request .md URLs (e.g. yoursite.com/my-post.md)
2. Plugin intercepts and identifies the bot
3. Converts your HTML to clean Markdown locally on your server
4. Caches the result for instant future delivery
5. Tracks the visit in your analytics dashboard

---

Features

Local HTML to Markdown Conversion

• Zero external API calls — everything runs on your server
• Uses league/html-to-markdown PHP library
• Response times of 1–4ms from cache
• YAML frontmatter with title, date, author, categories, tags
• Pre-generation option to warm cache before bots arrive

Bot Detection & Analytics

• Detects ChatGPT (GPTBot), Claude (ClaudeBot), Perplexity, Google Extended, Bing, Facebook, and Cohere bots
• Multi-layer detection pipeline: pattern matching + heuristic rules
• Real-time analytics dashboard — visits, trends, top pages
• Country tracking, IP geolocation, session tracking
• Browser & device tracking for citation clicks
• HTTP status code tracking to identify broken citations (404s, 500s)

LLM Traffic Dashboard

• See which AI platforms are citing your content
• Filter by date, platform, page, and request type
• CSV and JSON export
• Weekly and daily trend charts

Citation Alerts

• Get notified on citation from a AI platform

Multi-tier Caching

• Memory cache (static variables)
• Object cache (Redis/Memcached if available)
• Transient cache (database fallback)
• Tag-based cache invalidation
• Cache browser UI to view and manage cached entries

Headless WordPress Support

• AJAX-first architecture — works even when REST API is blocked by security plugins
• Auto-detects headless environments (Next.js, Gatsby, Nuxt)
• Client-side citation tracker (citation-tracker.js) captures real browser user agents
• Request type classification: html_page, rsc_prefetch, js_fallback, api_call

Google Analytics 4 Integration

• Sends bot visit data to GA4 via Measurement Protocol
• Connection testing from admin panel
• Optional — works without GA4

System Health & Auto-Repair

• Real-time health check dashboard
• Automatic database schema migration
• Daily health check cron with auto-repair
• Security plugin auto-whitelisting (Wordfence, Solid Security, Sucuri, AIOS)

---

Requirements

Requirement	Minimum	Recommended
PHP	7.4	8.0+
WordPress	5.8	Latest
PHP Extension	DOMDocument	(usually included by default)
Composer	Required for dev install	—

---

Installation

Option 1 — Upload ZIP (Easiest)

1. Download the latest release ZIP from the Releases page
2. Go to WordPress Admin → Plugins → Add New → Upload Plugin
3. Upload the ZIP and click Activate

Important: Always use the ZIP from the Releases page — not a ZIP exported from GitHub's Code button. GitHub exports add a wrapper folder that breaks WordPress's plugin installer. The release ZIP extracts correctly as third-audience/.

For troubleshooting and detailed setup see INSTALLATION.md.

Option 2 — Composer

cd wp-content/plugins/
git clone https://github.com/spcaeo/third-audience-wordpress-plugin.git third-audience
cd third-audience
composer install --no-dev

Then activate the plugin from WordPress Admin → Plugins.

Option 3 — Manual (Development)

cd wp-content/plugins/
git clone https://github.com/spcaeo/third-audience-wordpress-plugin.git third-audience
cd third-audience
composer install

After activation, visit Bot Analytics → System Health to verify everything is working correctly.

---

Configuration

Basic Setup

1. Activate the plugin
2. Go to WordPress Admin → Bot Analytics → Settings
3. Choose which post types to optimize (Posts, Pages, custom types)
4. Enable Discovery Tags — adds <link> tags so AI crawlers can find your .md URLs
5. Set Cache TTL (default: 24 hours)

Headless WordPress

If you run a headless frontend (Next.js, Gatsby, etc.):

1. Go to Bot Analytics → Settings → Headless Setup
2. The plugin auto-detects your environment
3. Add the citation tracker script to your frontend — see HEADLESS-SETUP.md

Google Analytics 4 (Optional)

1. Go to Bot Analytics → Settings → GA4 Integration
2. Enter your GA4 Property ID and Measurement Protocol Secret
3. Click Test Connection to verify

Email Digest (Optional)

1. Go to Bot Analytics → Email Digest
2. Configure SMTP settings and recipient email
3. Choose daily digest frequency

---

REST API

Namespace: third-audience/v1

Endpoint	Method	Auth	Description
/health	GET	Public	System health check
/analytics	POST	Nonce	Submit bot visit data
/diagnostics	GET	Admin only	Detailed system diagnostics

The plugin auto-falls back to AJAX endpoints if REST API is blocked by security plugins.

---

Hooks & Filters

Actions

// Fires after a bot visit is tracked
do_action( 'ta_bot_visit_tracked', $post_id, $bot_name, $traffic_type );

// Fires when cache is cleared for a post
do_action( 'ta_cache_cleared', $post_id );

// Fires on daily health check cron
do_action( 'ta_daily_health_check' );

// Fires on cache warm cron
do_action( 'ta_cache_warm_cron' );

Filters

// Filter the markdown content before it is cached and served
add_filter( 'ta_markdown_content', function( $markdown, $post_id ) {
    // Modify $markdown here
    return $markdown;
}, 10, 2 );

// Filter which post types are enabled for markdown serving
add_filter( 'ta_enabled_post_types', function( $post_types ) {
    $post_types[] = 'my_custom_type';
    return $post_types;
} );

// Filter the YAML frontmatter before it is added to markdown
add_filter( 'ta_frontmatter', function( $frontmatter, $post_id ) {
    $frontmatter['custom_field'] = get_post_meta( $post_id, 'my_field', true );
    return $frontmatter;
}, 10, 2 );

---

Database Tables

Table	Purpose
wp_ta_bot_analytics	All bot visits and citation tracking (main table)
wp_ta_citation_alerts	Citation alert history
wp_ta_bot_patterns	External bot pattern database
wp_ta_competitor_benchmarking	Competitor benchmark results
wp_ta_unknown_bots	Queue for unknown bot auto-learning

The plugin handles all database migrations automatically on activation and upgrade.

---

FAQ

Does this slow down my website for human visitors? No. The plugin only intercepts requests for .md URLs or requests with Accept: text/markdown headers. Human visitors are completely unaffected.

Does my content get sent to any external service? No. All HTML-to-Markdown conversion happens locally on your server using the league/html-to-markdown PHP library. Your content never leaves your server.

Does it work with security plugins like Wordfence or Solid Security? Yes. The plugin auto-detects blocked REST APIs and switches to AJAX endpoints automatically. It also whitelists itself in common security plugins on activation.

Does it work with headless WordPress (Next.js, Gatsby)? Yes. The headless setup includes a client-side JavaScript tracker and AJAX-first architecture that works regardless of how your frontend is built.

Which AI bots are detected? GPTBot (ChatGPT), ClaudeBot (Anthropic), PerplexityBot, Google Extended, Bingbot, FacebookBot, Cohere AI, and more via the heuristic detection pipeline.

Can I block specific bots? Yes. Go to Bot Analytics → Bot Management to block or allow specific bots granularly.

What happens when I uninstall the plugin? All plugin data is removed cleanly: database tables, options, transients, log files, and cron jobs are all deleted on uninstall.

---

Contributing

Contributions are welcome. To get started:

1. Fork the repository
2. Clone your fork locally
3. Install dependencies: composer install inside the project root
4. Make your changes on a new branch
5. Run tests: vendor/bin/phpunit inside the project root
6. Submit a pull request

Please follow WordPress coding standards. For major changes, open an issue first to discuss what you would like to change.

Running Tests

cd third-audience
vendor/bin/phpunit

Test files are in the tests/ directory. See TESTING-GUIDE.md for the full testing guide.

---

Changelog

See CHANGELOG.md for the full version history.

Latest: v3.6.1

• Bing fixed — plain Bing is now tracked as organic search, not "Bing AI"; only Bing Copilot counts as AI (with an automatic migration for old data)
• Hidden-referrer (Claude) detection on direct WordPress, plus scoped "Clear LLM Traffic / Bot Crawl" buttons
• OKF rebuilds in the background — post and menu saves no longer freeze on large sites

v3.6.0

• OKF (Open Knowledge Format) bundle served at /okf/ — a self-contained Markdown bundle of your whole site for AI agents
• .txt URL support — every page is also served at /<path>.txt (same content as .md, as text/plain)
• Redesigned LLM Traffic dashboard — drill-down cards (platform, country, browser, device, page type) → modal with the exact visits + CSV export, date-range filter
• Distinct platform colours, LLM-only Traffic Trend chart; About page documents .md / .txt / /okf/ discovery

v3.5.4

• Fixed fatal error on fresh WordPress installs — activation hook no longer blocks on HTTP requests
• Fixed SHOW COLUMNS FROM MySQL error when table doesn't exist yet on first install
• Fixed database auto-fixer creating wrong table schema on fresh installs
• Added INSTALLATION.md with full setup guide and troubleshooting

v3.5.3

• Dashboard enhancements, session dedup, headless data fixes
• Client user agent tracking for real browser data
• HTTP status code tracking to identify broken citations
• Request type classification (html_page, rsc_prefetch, js_fallback, api_call)

---

Credits

Author: Space-O Technologies

Support: Open a GitHub Issue

---

License

This plugin is licensed under the GNU General Public License v2.0 or later.

Copyright (c) 2026 Space-O Technologies

You are free to use, modify, and distribute this plugin under the terms of the GPL v2 license.