Love2Vec Documentation

A comprehensive guide to operating the Love2Vec mathematical topology engine.

Operating Philosophy: We do not match strings. We match semantics. Every tool in this platform relies on converting raw HTML into multi-dimensional space. To get accurate outputs, you must first feed the engine accurate inputs via the Data Ingestion tab.

1. Workspace Roster & Initialization

Before you can analyze a website, you must initialize a localized mathematical universe for it, known as a Workspace.

Creation: Click "New Workspace". The system will instantly provision a dedicated Vector Collection inside our Vector Database.
Open: Clicking this launches the 3D Physics engine. Note: If a workspace has 0 mapped URLs, the universe will be empty.

2. Data Ingestion (The AI Vectorizer)

This is the brain of the operation. You can feed the system either an XML Sitemap or a raw list of custom URLs.

How it works under the hood:

Sanitization: The crawler visits the page, brutally stripping out navigation menus, footers, scripts, and styling so the AI isn't confused by boilerplate UI.
Distillation: An advanced LLM reads the raw, cleaned text and distills the entire page into a hyper-granular topic.
Vectorization (Embeddings): That topic is converted into thousands of mathematical coordinates and mapped permanently into your Vector database.
Asynchronous Queue: To handle large sites, ingestion runs in the background using our robust Queue Manager and Daemon Workers, allowing you to queue thousands of URLs safely.
Topology Baking: Our systems automatically compress those thousands of dimensions down into an XYZ matrix using Principal Component Analysis (PCA) so human eyes can comprehend it.

3. 3D Latent Space Comparator & Visualizer

Traditional architecture maps look like flowcharts. Real search algorithms see architecture as a solar system of gravity and relationships. The Comparator visually proves how well your topical clusters are grouped.

Context Mode (PCA): Shows the mathematical reality of your data. Nodes that are physically closer together share a tighter semantic relationship. If your core pillar pages are scattered randomly across the matrix, search engines are confused by your structure.
Force Mode: Applies gravity based on internal PageRank and incoming links, showing you the "heaviest" pages on your site.
Crawl Tree Mode: Strips away cross-linking noise to show the strict top-down Spanning Tree hierarchy that bot crawlers experience.
Tooltips: Hover over any node to see the AI-generated distilled topic, proving exactly how the machine interprets your content.

4. Macro Views & High-Speed Snapshots

When you need to share your topology with stakeholders or monitor multiple workspaces, the engine provides high-performance rendering.

Macroverse Dashboard: An eagle-eye view of all your parsed sites in a single interface. Monitor network health, total vectors, and cluster density across your entire portfolio.
GZIP Snapshots: Once a topology matrix is baked, it can be exported as a highly compressed GZIP payload. This allows unauthenticated users or clients to view instantaneous, read-only 3D models via permanent shareable links without logging in or recalculating math on the fly.

5. Custom Metrics & Network Data

Topology isn't just about semantics; it's about business value and link equity.

Metrics Upload: Bind external third-party data (Ahrefs traffic, GA4 conversions, backlink counts) directly to your 3D nodes by uploading a CSV. The engine hashes the URLs and seamlessly merges your custom metrics into the graph space.
Network Metrics: Compare normalized internal link decay and structural graph density across different workspaces to see which site architectures distribute authority most efficiently.

6. Precision Tooling & Global Regex Filtering

Love2Vec features advanced analytical engines to audit your graph. Because large websites can contain thousands of URLs, analyzing the entire domain at once can create unnecessary noise.

The Regex (Regular Expression) Superpower: All advanced tools include Include Regex and Exclude Regex input fields to filter your URLs and surgically isolate your analysis.

📖 Read Regex Masterclass Guide

A. Thematic Gap Analysis

You cannot beat a competitor by copying their exact keywords. You beat them by identifying structural voids in their knowledge graph. The Thematic Gap tool performs a cross-dimensional vector search to find mathematical concepts your competitor possesses that you lack.

Semantic Clustering: The engine gathers all the missing semantic voids and runs them through a K-Means clustering algorithm, grouping missing articles into macro-themes.
LLM Naming: The AI evaluates each cluster and generates a "Silo Strategy" name so you know exactly what categories of content to commission next.

Regex Tip: Exclude /author/ or /tag/ directories so the AI doesn't flag missing paginated archive pages as "content gaps."

B. Content Cannibalization Matrix

If two URLs share a vector similarity that is mathematically too tight, they are competing for the same search intent and cannibalizing your ranking power.

Similarity Threshold: The tool scans your isolated URLs. Anything crossing your defined threshold (e.g., 90% match) is flagged as a conflict.
Merge Strategies: The engine automatically recommends whether to 301 redirect the weaker page, merge the content, or canonicalize based on PageRank and semantic density.

Regex Tip: Use the Include filter to isolate /services/. This ensures you only see cannibalization conflicts between your core landing pages, ignoring blog post overlap.

C. Hub & Spoke Architect

Search engines reward topical authority. This tool mathematically restructures your existing flat content into highly optimized Hub & Spoke silos.

Centroid Discovery: The algorithm groups your pages into K clusters, then mathematically calculates the exact center (centroid) of each cluster to determine which page deserves to be the "Hub" URL.
Spoke Mapping: All surrounding pages are mapped as "Spokes," complete with their Cosine Similarity score to the Hub, generating a perfect internal linking blueprint.

Regex Tip: Target a specific category by including /category-name/ to let the AI organize just that specific silo into a perfect Hub and Spoke structure.

D. Information Gain Analyzer

Google's Helpful Content system punishes websites that regurgitate standard facts without adding unique value. The Information Gain tool measures the true depth of your pages.

Density Scoring: Measures the semantic weight and unique vector footprint of a page compared to the rest of your site.
Thin Content Detection: Instantly isolates low-effort pages that need to be rewritten, expanded, or pruned to protect your domain-wide quality score.

Regex Tip: Analyze user-generated content or forum threads by including /community/ to find the highest-value discussions to promote.

E. Internal PageRank Sculptor

The Sculptor tool analyzes the physical link graph scraped during the ingestion phase to show you how link equity (PageRank) flows through your architecture.

Flow Visualization: Identifies "power nodes" hoarding link equity and "orphan nodes" that are starved of authority.
Optimization: Highlights exactly where you need to add strategic internal links to push ranking power to your money pages.

Regex Tip: Use the Regex Exclude field to filter out sitewide links like /privacy-policy or /contact so they don't skew your PageRank distribution metrics.

F. Taxonomy Architect & Keyword Miner

Move beyond flat lists with tools designed to restructure and extract maximum value from your text.

Taxonomy Architect: Synthesizes your current chaotic multi-level hierarchy into rigid, mathematically sound semantic taxonomy silos.
Keyword Miner: Deep-scans your processed payloads to extract high-value entities and clusters, offering a raw view into the exact phrases defining your matrix.

7. Workspace Maintenance & Post-Processing

Because vector math relies on the whole graph, updating one node can require recalculating the matrix. Our Maintenance tools keep your databases pristine.

Graph Repair: Clean up broken linkages, purge orphaned URLs, and fix database mismatches instantly.
Force Topology Bake: Manually trigger the PCA engine and K-Means algorithms to recalculate your XYZ coordinates if you have mass-deleted pages.
Data Inspector: Individually examine a URL to see the exact text extracted and the distilled topic generated by the LLM.

8. Settings, Integrations & Billing

Control your data pipelines and compute costs natively.

Integrations (BYOK): Bring Your Own Key. Plug in your own API keys for your preferred LLMs to bypass platform token costs completely. We do charge a % of usage to cover our costs.
Ledger: View real-time token usage, monitor your wallet balance, and top-up API credits securely via the integrated Payment gateway.

9. Exact Slug Match Migration

Designed for technical migrations. It bypasses the AI completely and compares the raw URL slugs between a baseline site and a target site to ensure no architecture was dropped during a rebuild.

10. Bulk Redirect Mapper

Site migrations usually require days of manual spreadsheet mapping. Our Redirect Mapper automates this using spatial mathematics.

Provide a list of "Old URLs". The engine vectorizes their slugs and queries your new Workspace. It returns the top 3 mathematically closest pages on your new site, complete with a confidence score. A score of 1.00 is a perfect semantic match. Anything below 0.75 suggests you deleted content during the migration without building a replacement for it.