AI Provider for Hugging Face vs HandL AI Connector Access Control

This plugin registers Hugging Face as an AI provider for the WordPress AI Client introduced in WordPress 7.0. It enables text and image generation through Hugging Face’s Inference API using popular open-source models. Features: Registers Hugging Face on the Settings > Connectors admin screen Supports API key management (environment variable, PHP constant, or database) Text generation via wp_ai_client_prompt() Image generation via wp_ai_client_prompt()->generate_image() Dynamically fetches top 20 popular models from HuggingFace (updated every 12 hours) Multi-provider support — automatically routes to the best available inference provider (hf-inference, fal-ai, replicate, together, nscale, wavespeed) Add custom models from the Settings > Hugging Face admin page Extensible model lists via hugging_face_ai_provider_models and hugging_face_ai_provider_image_models filters Configurable base URL for self-hosted TGI instances Clear, actionable error messages for common API issues Usage: $result = wp_ai_client_prompt( 'Summarize the benefits of caching.' ) ->using_temperature( 0.7 ) ->generate_text(); // Generate an image $image = wp_ai_client_prompt( 'A futuristic WordPress logo in neon colors' ) ->generate_image(); External Services This plugin connects to the Hugging Face Inference API to provide AI-powered text and image generation capabilities within WordPress. The Hugging Face Inference API is a third-party service operated by Hugging Face, Inc. What data is sent and when Your Hugging Face API key is sent with every request for authentication. When fetching the model list, a request is sent to the Hugging Face API to retrieve available models. When generating text, your prompt and any conversation context are sent to the Hugging Face Inference API. When generating images, your text prompt is sent to the Hugging Face Inference API. When resolving inference providers, a request is sent to the Hugging Face API to determine available providers for the selected model. Data is only sent when the plugin is actively used to generate text or images, or when loading the model list on the admin settings page. Service links Hugging Face Website Hugging Face Inference API Documentation Hugging Face Terms of Service Hugging Face Privacy Policy Hugging Face Content Policy All API communication is routed through https://router.huggingface.co and https://huggingface.co/api.

HandL AI Connector Access Control lets administrators allow/deny AI Client prompt execution on a per-plugin basis using the wp_ai_client_prevent_prompt filter introduced with the WordPress AI Client. Default behavior is allow. Learn mode (Audit & log tab) logs every AI Client call without blocking, so you can discover callers before enabling deny rules on the Plugin rules tab. Emergency kill switch blocks all AI Client calls except plugins you list as exceptions. Caller attribution is best-effort and is determined by inspecting the PHP call stack and mapping file paths to installed plugins. Privacy / Data This plugin does not send data to any external service. If you enable recent-call logging in Settings → HandL AI Connector Access Control, it stores a local log in the WordPress options table containing: Timestamp Allow/deny decision AI Client operation (e.g. generate_text, is_supported_for_text_generation) Provider and model when set on the prompt builder (or model preferences) Truncated prompt preview and selected generation config (best-effort) Input and output token counts when the AI Client completes a generation (best-effort) Best-effort calling plugin (plugin basename) and source file Current user id and display name Request URI Logs are kept as a count-based ring buffer (default 200 entries, configurable 20–1000). There is no time-based TTL—older rows drop only when the buffer is full. Logging is disabled by default.