Both make WordPress legible to AI search. LovedByAI serves an AI-only copy of your site from its own CDN; AI Data Layer builds a Wikidata-verified knowledge graph that lives on your own domain, and a live MCP server AI assistants can query directly.
Both plugins make your WordPress content easier for AI search engines to understand and cite, they just take opposite routes. LovedByAI generates a separate, AI-readable copy of your pages and serves it from its own CDN: your visitors see your site, the bots see LovedByAI’s version. AI Data Layer does the reverse, it publishes meaning on your own domain as structured data, clean markdown, machine-readable endpoints and a live MCP server, all grounded in a verified Wikidata knowledge graph. One rents you AI visibility; the other builds you an asset you own.
An honest side-by-side, including where LovedByAI does the job well.
| Capability | LovedByAI | AI Data Layer |
|---|---|---|
| Where your AI content lives | Separate AI copy on their CDN | Your own domain |
| Entity grounding | Not as verified IDs | Verified Wikidata QIDs + sameAs |
| Knowledge graph & entity registry | Not included | Site-wide graph + /entities/{id} |
| Topic clusters & cohesion score | Not included | 0–100 score |
| JSON-LD on your real pages | Applied to the AI copy | On your real pages |
| /llms.txt AI site index | Generated | On your domain |
| Per-page Markdown (.md) + llms-full.txt | Not included | Append .md to any URL |
| Public REST API (/wp-json/ai/v1/*) | Not included | Included |
| Built-in MCP server (agent-ready) | Not included | /wp-json/ai/v1/mcp |
| Human review & approve | Fully automated | Nothing publishes without sign-off |
| AI traffic / mentions / model attribution | Included | Not currently |
| Front-end performance impact | None (separate copy) | Negligible |
| SEO-plugin & WooCommerce aware | Works alongside | Defers to your SEO plugin |
| Setup | Plugin + free account | One-click + free trial, no card |
| Pricing | Free tier, then from $25/mo | Free trial, then from £12/mo |
LovedByAI optimises a copy of your site on its CDN. AI Data Layer publishes on the domain you own.
LovedByAI’s optimisations apply to a separate AI version of your pages, hosted on LovedByAI’s CDN and served to AI crawlers, a clever, zero-touch model, but your AI visibility lives on infrastructure you don’t control, and cancelling takes the optimised copy with it. AI Data Layer publishes everything on your own site: JSON-LD injected into your real pages, an /llms.txt index on your domain, a clean markdown version of every page (just append .md to any URL), public REST endpoints, and your own MCP server. It’s a permanent asset on a domain you own, not a copy you rent.
AI Data Layer grounds every entity to a real Wikidata ID; LovedByAI reformats content page by page.
LovedByAI focuses on reformatting and structuring your content for AI, schema detection, AI-query formatting, FAQ generation. AI Data Layer builds something different: a verified knowledge graph. It extracts 20–40+ entities per page, grounds each against Wikidata with a real QID and sameAs links only when confidently matched, and drops uncertain matches rather than guessing, so AI systems never get fed a “company X is actually a biomarker” error. Those entities are reconciled site-wide, stored in a deduplicated registry, and published as AI-readable reference pages at /entities/{id}.
Every AI Data Layer site becomes its own MCP server AI assistants can query as a tool.
This is the one no competitor in the AI-search-for-WordPress space currently matches. Every site running AI Data Layer becomes its own MCP (Model Context Protocol) server at /wp-json/ai/v1/mcp, with zero extra hosting. Assistants like Claude, ChatGPT custom connectors, Cursor and Perplexity can connect directly and query your knowledge, search, fetch, get_page_context, list_entities, get_entity, get_site_overview. LovedByAI publishes static files for crawlers to find; AI Data Layer does that and exposes a live, queryable tool server.
AI Data Layer shows every suggestion for human approval; LovedByAI runs fully automated.
LovedByAI implements everything automatically, for some teams that hands-off model is exactly the appeal. AI Data Layer takes the opposite stance by design: every AI suggestion (entities, schema, FAQs) is shown for human approval with inline editing before it goes live, and related-page suggestions only ever point to real pages that share a grounded entity, never to AI-invented links. If you care about editorial integrity and accuracy over pure automation, that control matters.