rsnare

dAIrect Chat enables you to embed a Gemini-powered chatbot into any page or post using a shortcode. Designed for simplicity, flexibility, and fast response, it is perfect for customer support, creative interaction, or building your own branded assistant.
<h4>Features include:</h4>
<ul>
<li>Multi-turn memory (stored per user)</li>
<li>Configurable: temperature, topP, max tokens, timeout</li>
<li>Custom system instruction (acts as a personality or role)</li>
<li>Voice input (Chrome only)</li>
<li>Easy shortcode integration: <code>[dairect_chat_widget]</code></li>
<li>Admin test prompt with saved parameter preview</li>
</ul>
<h3>External Services</h3>
This plugin connects to the Google Gemini API to generate conversational responses using AI.
<ul>
<li>Data sent: Chat message contents and optional system instructions.</li>
<li>Sent when: The user sends a message via the widget or admin test prompt.</li>
<li>Service provider: Google LLC 
<a href="https://policies.google.com/terms" rel="nofollow ugc">Terms of Service</a> | <a href="https://policies.google.com/privacy" rel="nofollow ugc">Privacy Policy</a></li>
</ul>
<h3>License</h3>
GPLv2 or later — Fork freely, build professionally.
License URI: https://www.gnu.org/licenses/gpl-2.0.html
MIT — Fork freely, build professionally.

dAIrect Chat

dAIrect Chat is a lightweight Gemini AI chat widget for WordPress with multi-turn memory, voice input, and custom system instructions.

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.

dAIrect Chat vs Ultimate AI Connector for Compatible Endpoints