The WordPress.org plugin repository contains roughly 212 plugins that mention AI in their description as of June 2026. A year ago that number was closer to 47. Growth is real, but the signal-to-noise has dropped faster than the count rose: about 30 of those plugins are AI-native (built for LLM integration from the ground up), the other ~182 are existing plugins that bolted a “generate” button onto established product lines.
The honest map of where WordPress + AI actually stands looks like this.
The natives
Five plugins drive the AI-native conversation:
- AI Engine (Meow Apps). Most-featured by surface area. Multi-provider out of the box, MCP server endpoint, chatbot, content generation, image generation. Run solo by Jordy Meow since 2022. Genuinely usable free tier.
- AI Power (10Web). Hosting-vendor-backed, integrated tightly with the 10Web stack but works standalone. Strong on landing-page generation.
- GetGenie. Content-marketing focused. SEO-first prompts, internal linking suggestions, competitor analysis.
- Bertha AI. Older entrant, lost some momentum after the 2023 funding crunch. Still shipping, slower cadence.
- AIKit. Newer, growing. Vibes-based content generation with a chat-style interface.
The natives compete on two axes: feature surface (how many distinct things can the plugin do) and provider breadth (how many LLMs does it support out of the box). AI Engine leads on both. Others differentiate on niche use cases.
The bolt-ons
Existing plugins added AI features in the last 18 months. The volume leaders:
- Yoast SEO — title and meta generation, content improvement suggestions, internal linking
- Elementor — image and text generation inside the page builder, integrated as Elementor AI
- WPForms, Gravity Forms — form prompt generation, conditional logic suggestions
- WooCommerce — product description generation, category copy
The bolt-ons benefit from existing distribution. Each one ships AI features to millions of sites overnight, just by updating. None of them are AI-first products, but they add up to a lot of WordPress AI use in aggregate. Most of the actual AI calls happening from WordPress sites today come from bolt-on plugins, not from the AI-natives.
Provider split
Roughly 163 plugins integrate OpenAI. About 28 integrate Anthropic. Around 21 integrate Google Gemini. Multi-provider plugins exist (AI Engine being the most prominent) but most plugins are single-provider, and OpenAI is the default.
Anthropic share is growing fast as Claude 4 outpaces GPT-5 on the kind of structured reasoning and code-heavy tasks that WordPress workflows favor (plugin code review, migration scripts, complex schema generation). Google share remains flat outside of plugins specifically targeting Google Ads or YouTube creators.
The MCP frontier
Model Context Protocol launched in late 2024 and has now started to land in WordPress. As of June 2026, the MCP servers for WordPress count is small:
- AI Engine’s built-in MCP endpoint (Pro tier). Authenticated, exposes post management, media library, and a few WP-CLI bridges.
- WP-MCP plugin by an independent two-person team. Open-source, narrower feature surface, but cleaner protocol implementation.
- WP-CLI MCP bridge in early development. Exposes WP-CLI commands as MCP tools.
- A community Cloudflare Worker that mediates MCP traffic and signs requests on behalf of an authenticated agent.
Six months from now, expect at least one of these to become the canonical choice. The first vendor that ships a “let your AI agent maintain your WordPress site” experience that genuinely works will win an outsized share of the agentic workflow market that’s about to open up.
The two ways this plays out: either a major hosting vendor (Hostinger, WP Engine, Kinsta) ships a managed MCP-for-WP service and commoditizes it, or one of the AI-native plugins matures their MCP server to the point that it becomes the de facto standard. Either way, the next twelve months matter.
What’s still missing
- RAG over your own content. A handful of plugins do early-stage retrieval-augmented generation over the WordPress database, but none are production-ready. The use case is screaming for serious work: a chatbot that knows your entire site’s content with proper citation.
- Genuine agentic workflows. An AI that maintains the site (handles plugin updates, watches for security advisories, optimizes images on upload, triages support tickets). Several teams claim to be building this. None have shipped it.
- Compliance tooling. EU AI Act took effect last summer. WordPress sites that publish AI-generated content fall under transparency requirements depending on jurisdiction and content type. No plugin in the repo handles this well yet, and the legal risk is non-zero.
- AI-assisted plugin development from inside WP admin. WP-CLI + Claude + your own context = workflow that some agencies are running manually. Nobody has packaged it into a plugin.
The half-baked obvious thing
Nobody has shipped a polished AI image alt-text generator that runs automatically on upload, with cost controls and an audit log. Several plugins technically do this but none of them do it well. The market is sitting there waiting for someone to build the thoughtful version with rate limiting, vision model fallbacks, a useful admin review queue, and a budget cap.
If you’re an indie plugin developer reading this, that’s a category with one good launch away from being the default choice.
Six-month outlook
What to expect by year-end 2026:
- One MCP-for-WordPress server becomes the canonical default
- AI page builders consolidate: Elementor AI matures, Divi AI ships a second-tier offering, Bricks AI emerges as a real third option
- A hosted “Cloudflare AI Gateway for WordPress” or equivalent emerges
- EU AI Act compliance becomes a real plugin category
- At least one agentic workflow plugin ships that genuinely works for narrow but useful tasks (support triage, content moderation)
The call
If you’re picking an AI plugin for a WordPress site right now, the default safe choice is AI Engine (paired with your own LLM provider keys). The breadth of features, the active development, and the early MCP support all make it the strongest single-plugin bet for the next twelve months.
If you’re building a WordPress site for a client and the AI features matter, assume the bolt-on AI tools shipped by Yoast and Elementor will be invisible defaults. You don’t need to install anything extra. The interesting work is everywhere else.
If you’re a developer or agency looking at where to invest learning time over the next quarter, the MCP frontier is the place. The plugin natives are mature enough that the marginal value of learning a new one is low; the canonical MCP-for-WP doesn’t exist yet, and being early matters.
Methodology
Plugin counts derived from WordPress.org repository search results in May 2026, manually deduplicated for plugins that mention AI but don’t ship AI features. Provider integration counts approximate, derived from plugin readme files, feature pages, and changelog text. Specific plugins called out are tracked in the Plugins directory and (when AI-native) in the AI Plugins tracker.
This piece reflects mid-2026 reality. The landscape will change. We’ll publish the next state of in Q4 2026 with updated counts and revised calls.