David Stone

This plugin extends the WordPress AI Client to support any AI service or server that uses the standard chat completions API format (<code>/v1/chat/completions</code> and <code>/v1/models</code> endpoints).
Supported services include:
<ul>
<li>Ollama – Run open-source models (Llama, Mistral, Gemma, etc.) locally on your own hardware.</li>
<li>LM Studio – Desktop application for local LLM inference with a one-click server.</li>
<li>OpenRouter – Unified API providing access to 100+ models from multiple providers.</li>
<li>vLLM – High-throughput inference server for production deployments.</li>
<li>LocalAI – Drop-in replacement for running models locally.</li>
<li>text-generation-webui – Popular web UI with API server mode.</li>
<li>Any compatible endpoint – Works with any server implementing the standard format.</li>
</ul>
Requirements:
<ul>
<li>WordPress 7.0+ – The AI Client SDK is included in core. This plugin works on its own without any additional dependencies.</li>
<li>WordPress 6.9 – Also supported when the Gutenberg plugin (23.0+) is active, which provides the AI Client SDK.</li>
</ul>
Why it matters:
Other AI-powered plugins that use the WordPress AI Client (such as AI Experiments) can automatically discover and use any model you connect through this plugin. Configure your endpoint once and every AI feature on your site can use it.
How it works:
<ol>
<li>Install and activate the plugin.</li>
<li>Go to Settings > Connectors and configure the connector with your endpoint URL (e.g. <code>http://localhost:11434/v1</code> for Ollama).</li>
<li>Optionally provide an API key for services that require authentication.</li>
<li>The plugin registers a provider with the WordPress AI Client and dynamically discovers all available models from your endpoint.</li>
</ol>
The plugin also handles practical concerns like extended HTTP timeouts for slow local inference and non-standard port support.

Ultimate AI Connector for Compatible Endpoints

Connects the WordPress AI Client to Ollama, LM Studio, or any AI endpoint that uses the standard chat completions API format.

AcrossWP

This plugin provides llama.cpp integration for the WordPress AI Client. It lets WordPress sites use large language models running via a llama.cpp server for text generation and other AI capabilities.
llama.cpp exposes an OpenAI-compatible API, and this provider uses that API to communicate with any GGUF model loaded into your llama.cpp server.
Features:
<ul>
<li>Text generation with any llama.cpp-loaded model</li>
<li>Automatic model discovery from your llama.cpp server</li>
<li>Function calling support</li>
<li>Structured output (JSON mode) support</li>
<li>Settings page to configure the server URL (default: http://127.0.0.1:8080)</li>
<li>Works without an API key for local instances</li>
</ul>
Requirements:
<ul>
<li>PHP 7.4 or higher</li>
<li>WordPress AI Client plugin must be installed and activated</li>
<li>llama.cpp server running locally or on a remote host</li>
</ul>
<h3>Getting Started</h3>
<h4>What do I need before using this plugin?</h4>
You need a running llama.cpp server (local or remote) and the WordPress AI Client plugin installed and activated.
<h4>How do I install llama.cpp?</h4>
On macOS run <code>brew install llama.cpp</code>. For other platforms, see the official docs at https://llama-cpp.com/download/
<h4>Where do I get a model?</h4>
Download a GGUF model from Hugging Face. TinyLlama 1.1B (Q4_K_M, ~636 MB) is a good starting point for testing.
<h4>How do I start the server?</h4>
Run <code>llama-server --models-dir ~/models</code>. The server starts on <code>http://127.0.0.1:8080</code> by default.
<h4>How do I connect WordPress to my llama.cpp server?</h4>
Go to Settings > llama.cpp and enter your server URL. Leave it blank to use the default (<code>http://127.0.0.1:8080</code>).
For full step-by-step setup instructions, see the <a href="https://github.com/acrosswp/ai-provider-for-llamacpp/" rel="nofollow ugc">Getting Started Guide</a>.

AI Provider for llama.cpp

llama.cpp provider for the WordPress AI Client.

Ultimate AI Connector for Compatible Endpoints vs AI Provider for llama.cpp