mittwald

LukasFritzeDev

Martin Helmich

This plugin integrates <a href="https://developer.mittwald.de/docs/v2/platform/aihosting/" rel="nofollow ugc">mittwald AI Hosting</a> with WordPress AI features, enabling AI-powered features on your WordPress site using mittwald’s infrastructure.
This plugin requires WordPress 7.0 or newer.
<h3>Supported Operations & Models</h3>
<h4>Chat Completions</h4>
Fully supported for conversational AI, content generation, and chat-based interactions.
Available Models: 
– GPT-OSS models: Open-source GPT-compatible models 
– Qwen: Qwen 3.5 122B-FP8 and Qwen 3.6 35B-FP8 (supports vision/image input) 
– Ministral: supports vision/image input 
– Devstral: optimized for code generation
Capabilities: 
– Standard text chat 
– Image vision (Mistral Small and Qwen models) 
– JSON output formatting 
– Tool/function calling 
– Streaming responses

AI Provider for mittwald

Connects WordPress AI to mittwald AI Hosting, enabling AI-powered features via an OpenAI-compatible provider integration.

Per Søderlind

This plugin registers exo as an AI provider in WordPress 7’s AI Client SDK and Connectors page.
<a href="https://github.com/exo-explore/exo" rel="nofollow ugc">exo</a> connects all your devices into an AI cluster, enabling you to run frontier models locally. It exposes an OpenAI-compatible API that this plugin connects to.
Features:
<ul>
<li>Registers exo as a WordPress AI provider</li>
<li>OpenAI-compatible text generation via exo’s chat completions API</li>
<li>Auto-detect active models from your running exo cluster</li>
<li>Capability detection — displays model capabilities (Text, Code, Thinking, Vision) as badges</li>
<li>“Connect & Detect” / “Save & Re-detect” connector flow</li>
<li>Optional API key authentication with secure storage</li>
<li>Configurable endpoint URL (default: <code>http://localhost:52415</code>)</li>
<li>Settings integrated into WordPress 7’s Connectors page</li>
</ul>
Choosing a Model:
exo exposes every model in its catalog, but only models actively loaded on your cluster will respond. Use “Connect & Detect” on the Connectors page to discover which models are running.
Recommended — Instruct models produce clean, usable output for WordPress AI features (title generation, content suggestions, etc.):
<ul>
<li><code>Llama-3.2-3B-Instruct-8bit</code> — ~3 GB, fast, great for short tasks</li>
<li><code>Meta-Llama-3.1-8B-Instruct-4bit</code> — ~4 GB, good balance of speed and quality</li>
<li><code>Llama-3.3-70B-Instruct-4bit</code> — ~35 GB, best quality, needs a larger cluster</li>
</ul>
Avoid — Reasoning/thinking models (Qwen3.5, DeepSeek, GLM, Nemotron-Nano) spend most tokens on internal chain-of-thought, producing slow responses with minimal visible output.
To load a model: <code>exo run mlx-community/Llama-3.2-3B-Instruct-8bit</code>
Requirements:
<ul>
<li>WordPress 7.0 or later</li>
<li>PHP 8.3 or later</li>
<li>A running exo cluster (see <a href="https://github.com/exo-explore/exo" rel="nofollow ugc">exo documentation</a>)</li>
</ul>
<h3>Configuration</h3>
The plugin can be configured via the Connectors page or environment variables:
<ul>
<li><code>AIPRFOEX_ENDPOINT</code> — exo API endpoint (default: <code>http://localhost:52415</code>)</li>
<li><code>AIPRFOEX_API_KEY</code> — Optional API key for authentication</li>
<li><code>AIPRFOEX_MODEL</code> — Model name to use (auto-detected if empty)</li>
</ul>
You can also define these as constants in <code>wp-config.php</code>.

AI Provider for exo

Connect WordPress to exo — run frontier AI models locally on your device cluster.

AI Provider for mittwald vs AI Provider for exo