Documentation

RelatIQ Vector Engine

Welcome to the future of content discovery. RelatIQ is a privacy-first, zero-tracking related posts plugin powered by a local Semantic Vector Engine. Unlike legacy plugins that rely on literal tags or expensive third-party cloud APIs, RelatIQ understands the contextual meaning of your content entirely on your own server.

1. System Requirements

To ensure the vector engine runs smoothly without impacting server performance, your environment should meet the following minimums:

WordPress

5.8 or higher

PHP

7.4+ (8.0+ highly recommended for math speed)

Database

MySQL 5.7+ or MariaDB 10.3+

Cache

Redis or Memcached object caching enabled

2. Installation & Quick Start

  • 1
    Install

    Upload the relatiq folder to /wp-content/plugins/, or upload the .zip file via Plugins > Add New in your dashboard.

  • 2
    Activate

    Activate the plugin from the plugin list.

  • 3
    Onboarding

    Upon activation, you will be redirected to the Magic Console Setup Wizard. Select the post types you want the AI to read, choose a layout, and let the system initialize.

  • 4
    Rebuild Vectors

    If you skipped the wizard, navigate to “RelatIQ Tools” and click “Start Background Rebuild”. The AI will safely read and tokenize existing content in micro-batches.

3. General & Display Settings

Navigate to RelatIQ > General & Display to customize the frontend appearance. The interface features a real-time Live Studio Preview to see your changes instantly.

Layout Types

RelatIQ includes six 2030-Era layouts, all wrapped in aggressive CSS shields to prevent your theme from breaking their typography:

  • Card
    Standard grid with image, title, meta, and excerpt.
  • List
    A compact, vertical list with small thumbnail masks.
  • Carousel
    A horizontal scrolling layout perfect for mobile devices.
  • Hero
    Features the most relevant post as a massive hero block.
  • Magazine
    A clean, horizontal split layout for news sites.
  • Minimal
    Text-only grid for ultra-fast load times.

Dynamic Controls & Injection

  • Excerpt Metric
    Choose between counting strictly by Words or by Characters.
  • Excerpt Length
    Use the interactive slider to set the exact cutoff length.

Note: Excerpts are sanitized to strip rogue shortcodes and HTML, guaranteeing they never break your layout.

Auto Injection (ON)
: Appends block to bottom of content.
Auto Injection (OFF)
: Use shortcode [rltiq_related_posts_ai]

4. Semantic Logic & Tuning

Navigate to RelatIQ > Semantic Logic & Tuning to adjust the “brain” of the plugin.

  • What to Analyze?: Select which post types the AI should read.
  • Strictness Threshold:
    0.1 = Loose. Yields more results, broad connections.
    0.5 = Balanced similarity.
    0.8 = Strict. Yields fewer results, highly identical topics.

  • Hide if No Match: Hides the widget entirely rather than showing fallback posts.
  • Exclude URLs & IDs: Blacklist specific Post IDs from being recommended.

5. Relations Manager

Navigate to RelatIQ > Relations Manager to audit your database table and see exactly what the AI has processed.

Vector Status

Ready
Missing

Cache Status

Active
Warning

Manual Actions

Schedule or Force Instant recalculations

6. Analytics & Reporting

Navigate to RelatIQ > Analytics & Reporting to view your performance data. We use a 2kb, zero-dependency Vanilla JS beacon to track clicks without slowing down your site or violating privacy laws.

Destination

Top Performing: Which recommended posts get clicked the most.

Origin

Top Source Pages: Which articles are best at keeping users reading.

7. Enterprise Tools & WP-CLI

RelatIQ is built for massive architectures (30,000+ posts) and WPMU Multisite networks.

Background Processing (Web)

In RelatIQ > Tools, you can trigger manual rebuilds. The system utilizes Document Frequency (DF) Capping and tiny micro-batches (25 posts at a time) to prevent InnoDB gap locks.

WP-CLI Support (Terminal)

For massive enterprise sites, we recommend bypassing the web server entirely using SSH:

wp relatiq rebuild_vectors --batch-size=500
# Rebuild all vectors in batches of 500
wp relatiq clear_cache
# Purge the HTML fragment cache
wp relatiq cleanup_db
# Clean up orphaned meta data

8. Frequently Asked Questions

Does RelatIQ use the OpenAI API or any cloud processing?
No. RelatIQ is 100% self-contained. Your content never leaves your server. No API keys, no monthly cloud fees, and zero external tracking.
Will the Vector Engine bloat my database?
No. RelatIQ completely bypasses the standard post meta table for its heavy lifting. It uses dedicated, composite-indexed custom tables designed for infinite scaling.
Does it support Object Caching?
Yes. If RelatIQ detects Redis or Memcached, it automatically shifts from database-cached fragments to RAM-based delivery, resulting in sub-millisecond rendering on the frontend.
My theme has very aggressive CSS. Will it break the layouts?
No. RelatIQ 3.0.0 utilizes aggressive inline CSS shields and Base64 SVG masks. Your theme cannot hijack our hyperlink colors or break the DOM structure of our grid layouts.
Does it support non-English languages?
Yes. RelatIQ includes a native Bigram Tokenizer specifically engineered to support CJK (Chinese, Japanese, Korean) languages, alongside standard space-based languages like English, Spanish, and French.