[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fHrZ6cRDgW05gzEVqeWi1HurXpTEuTSqkhmbcufb5dE0":3,"$fPPsp_HSYyfTGOsbCgnklgORginhiOSwHDTvtsH1ssaU":117,"$fqMq7gzBZ__mOr-2h8-dp1MSEkcgZXvLohNeMfXHVMp8":122},{"slug":4,"name":5,"version":6,"author":7,"author_profile":8,"description":9,"short_description":10,"active_installs":11,"downloaded":12,"rating":11,"num_ratings":11,"last_updated":13,"tested_up_to":14,"requires_at_least":15,"requires_php":16,"tags":17,"homepage":23,"download_link":24,"security_score":25,"vuln_count":11,"unpatched_count":11,"last_vuln_date":26,"fetched_at":27,"discovery_status":28,"vulnerabilities":29,"developer":30,"crawl_stats":26,"alternatives":38,"analysis":26,"fingerprints":26},"ai-provider-for-openai-compatible-servers","AI Provider for OpenAI Compatible Servers","1.0.0","George Stephanis","https:\u002F\u002Fprofiles.wordpress.org\u002Fgeorgestephanis\u002F","\u003Cp>This plugin registers a provider for the WordPress AI Client that talks to any server implementing the OpenAI Chat Completions API — including Ollama, LM Studio, vLLM, llama.cpp, and LocalAI.\u003C\u002Fp>\n\u003Cp>Run AI features in WordPress against models on your own hardware. No third-party AI service, no per-token billing, and no content leaving your network.\u003C\u002Fp>\n\u003Ch4>Features\u003C\u002Fh4>\n\u003Cul>\n\u003Cli>Connect to any OpenAI-compatible base URL, local or remote (defaults to Ollama’s \u003Ccode>http:\u002F\u002Flocalhost:11434\u002Fv1\u003C\u002Fcode>).\u003C\u002Fli>\n\u003Cli>Autodetect available models from the server’s \u003Ccode>\u002Fmodels\u003C\u002Fcode> endpoint, or specify models manually.\u003C\u002Fli>\n\u003Cli>Setup card on the Connectors settings screen with a one-click connection test.\u003C\u002Fli>\n\u003Cli>Best-effort detection of the running backend (Ollama, vLLM, or LM Studio), swapping in the brand’s name and icon on the connector card once detected.\u003C\u002Fli>\n\u003Cli>Optional API key (bearer token) for servers that require authentication.\u003C\u002Fli>\n\u003Cli>Custom HTTP headers for reverse proxies and gateways.\u003C\u002Fli>\n\u003Cli>Advanced options for reasoning models: disable thinking blocks, and an R1-style message format that folds the system prompt into the first user message (needed for DeepSeek-R1).\u003C\u002Fli>\n\u003Cli>Context length override and a vision\u002Fmultimodal support toggle.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch4>Requirements\u003C\u002Fh4>\n\u003Cul>\n\u003Cli>The WordPress AI Client must be available (WordPress 7.0+).\u003C\u002Fli>\n\u003Cli>A running OpenAI-compatible inference server reachable from your web server.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3>Developer Hooks\u003C\u002Fh3>\n\u003Cp>This plugin provides several filters for other plugins to dynamically extend its functionality:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>\u003Ccode>connectors_ai_openai_compatible_servers_request_data\u003C\u002Fcode>\u003C\u002Fstrong>: Filters the completions request body payload before it is sent to the endpoint. Passed arguments: \u003Ccode>$data\u003C\u002Fcode> (array), \u003Ccode>$modelId\u003C\u002Fcode> (string).\u003C\u002Fli>\n\u003Cli>\u003Cstrong>\u003Ccode>connectors_ai_openai_compatible_servers_request_headers\u003C\u002Fcode>\u003C\u002Fstrong>: Filters HTTP headers for all request pathways. Passed arguments: \u003Ccode>$headers\u003C\u002Fcode> (array), \u003Ccode>$path\u003C\u002Fcode> (string), \u003Ccode>$method\u003C\u002Fcode> (string).\u003C\u002Fli>\n\u003Cli>\u003Cstrong>\u003Ccode>connectors_ai_openai_compatible_servers_detect_provider\u003C\u002Fcode>\u003C\u002Fstrong>: Adds or overrides detection checks when probing server origins. Passed arguments: \u003Ccode>$detected_type\u003C\u002Fcode> (null|string), \u003Ccode>$base_url\u003C\u002Fcode> (string), \u003Ccode>$headers\u003C\u002Fcode> (array), \u003Ccode>$probe_args\u003C\u002Fcode> (array).\u003C\u002Fli>\n\u003Cli>\u003Cstrong>\u003Ccode>connectors_ai_openai_compatible_servers_model_options\u003C\u002Fcode>\u003C\u002Fstrong>: Modifies the \u003Ccode>SupportedOption\u003C\u002Fcode> array for a resolved model. Passed arguments: \u003Ccode>$options\u003C\u002Fcode> (array), \u003Ccode>$modelId\u003C\u002Fcode> (string), \u003Ccode>$mode\u003C\u002Fcode> (string: ‘manual’ or ‘autodetect’).\u003C\u002Fli>\n\u003Cli>\u003Cstrong>\u003Ccode>connectors_ai_openai_compatible_servers_model_capabilities\u003C\u002Fcode>\u003C\u002Fstrong>: Modifies the \u003Ccode>CapabilityEnum\u003C\u002Fcode> array for a resolved model. Passed arguments: \u003Ccode>$capabilities\u003C\u002Fcode> (array), \u003Ccode>$modelId\u003C\u002Fcode> (string), \u003Ccode>$mode\u003C\u002Fcode> (string: ‘manual’ or ‘autodetect’).\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3>Credits\u003C\u002Fh3>\n\u003Cp>The Ollama, vLLM, and LM Studio brand icons shown when a server is detected are from\u003Cbr \u002F>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flobehub\u002Flobe-icons\" rel=\"nofollow ugc\">LobeHub’s lobe-icons\u003C\u002Fa>, used under the MIT license.\u003C\u002Fp>\n","Connect local or self-hosted OpenAI-compatible servers (Ollama, LM Studio, vLLM, and more) to the WordPress AI Client.",0,68,"2026-07-18T22:11:00.000Z","7.0.2","7.0","7.4",[18,19,20,21,22],"ai","llm","local-ai","ollama","openai","https:\u002F\u002Fgithub.com\u002Fgeorgestephanis\u002Fai-provider-for-openai-compatible-servers","https:\u002F\u002Fdownloads.wordpress.org\u002Fplugin\u002Fai-provider-for-openai-compatible-servers.1.0.0.zip",100,null,"2026-07-22T17:31:50.256Z","no_bundle",[],{"slug":31,"display_name":7,"profile_url":8,"plugin_count":32,"total_installs":33,"avg_security_score":34,"avg_patch_time_days":35,"trust_score":36,"computed_at":37},"georgestephanis",19,15570,91,30,88,"2026-08-28T20:47:31.475Z",[39,55,70,89,103],{"slug":40,"name":41,"version":42,"author":43,"author_profile":44,"description":45,"short_description":46,"active_installs":47,"downloaded":48,"rating":25,"num_ratings":49,"last_updated":50,"tested_up_to":14,"requires_at_least":15,"requires_php":16,"tags":51,"homepage":53,"download_link":54,"security_score":25,"vuln_count":11,"unpatched_count":11,"last_vuln_date":26,"fetched_at":27},"ai-provider-for-ollama","AI Provider for Ollama","1.1.1","Fueled","https:\u002F\u002Fprofiles.wordpress.org\u002Ffueled\u002F","\u003Cp>This plugin provides \u003Ca href=\"https:\u002F\u002Follama.com\u002F\" rel=\"nofollow ugc\">Ollama\u003C\u002Fa> integration for the WordPress AI Client. It lets WordPress sites use large language models running locally or on a remote Ollama instance for text and image generation and other AI capabilities.\u003C\u002Fp>\n\u003Cp>Ollama exposes an \u003Ca href=\"https:\u002F\u002Follama.com\u002Fblog\u002Fopenai-compatibility\" rel=\"nofollow ugc\">OpenAI-compatible API\u003C\u002Fa>, and this provider uses that API to communicate with any model you have pulled into Ollama (Llama, Mistral, Gemma, Phi, and many more).\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Features:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Text generation with any Ollama model\u003C\u002Fli>\n\u003Cli>Image generation with supported models\u003C\u002Fli>\n\u003Cli>Automatic model discovery from your Ollama instance\u003C\u002Fli>\n\u003Cli>Function calling support\u003C\u002Fli>\n\u003Cli>Structured output (JSON mode) support\u003C\u002Fli>\n\u003Cli>Settings page for host URL and seeing available models\u003C\u002Fli>\n\u003Cli>Works without an API key for local instances\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cstrong>Requirements:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>PHP 7.4 or higher\u003C\u002Fli>\n\u003Cli>WordPress 7.0 or higher\u003C\u002Fli>\n\u003Cli>Ollama running locally or on a remote server (like Ollama Cloud)\u003C\u002Fli>\n\u003C\u002Ful>\n","Ollama provider for the WordPress AI Client.",200,1736,2,"2026-05-14T17:08:00.000Z",[18,52,19,20,21],"connector","https:\u002F\u002Fgithub.com\u002Ffueled\u002Fai-provider-for-ollama","https:\u002F\u002Fdownloads.wordpress.org\u002Fplugin\u002Fai-provider-for-ollama.1.1.1.zip",{"slug":56,"name":57,"version":58,"author":59,"author_profile":60,"description":61,"short_description":62,"active_installs":63,"downloaded":64,"rating":11,"num_ratings":11,"last_updated":65,"tested_up_to":14,"requires_at_least":66,"requires_php":16,"tags":67,"homepage":68,"download_link":69,"security_score":25,"vuln_count":11,"unpatched_count":11,"last_vuln_date":26,"fetched_at":27},"ultimate-ai-connector-compatible-endpoints","Ultimate AI Connector for Compatible Endpoints","2.1.2","David Stone","https:\u002F\u002Fprofiles.wordpress.org\u002Fsuperdav42\u002F","\u003Cp>This plugin extends the WordPress AI Client to support \u003Cstrong>any AI service or server that uses the standard chat completions API format\u003C\u002Fstrong> (\u003Ccode>\u002Fv1\u002Fchat\u002Fcompletions\u003C\u002Fcode> and \u003Ccode>\u002Fv1\u002Fmodels\u003C\u002Fcode> endpoints).\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Supported services include:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Ollama\u003C\u002Fstrong> – Run open-source models (Llama, Mistral, Gemma, etc.) locally on your own hardware.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>LM Studio\u003C\u002Fstrong> – Desktop application for local LLM inference with a one-click server.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>OpenRouter\u003C\u002Fstrong> – Unified API providing access to 100+ models from multiple providers.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>vLLM\u003C\u002Fstrong> – High-throughput inference server for production deployments.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>LocalAI\u003C\u002Fstrong> – Drop-in replacement for running models locally.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>text-generation-webui\u003C\u002Fstrong> – Popular web UI with API server mode.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Any compatible endpoint\u003C\u002Fstrong> – Works with any server implementing the standard format.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cstrong>Requirements:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>WordPress 7.0+\u003C\u002Fstrong> – The AI Client SDK is included in core. This plugin works on its own without any additional dependencies.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>WordPress 6.9\u003C\u002Fstrong> – Also supported when the Gutenberg plugin (23.0+) is active, which provides the AI Client SDK.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cstrong>Why it matters:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>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.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>How it works:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Col>\n\u003Cli>Install and activate the plugin.\u003C\u002Fli>\n\u003Cli>Go to \u003Cstrong>Settings > Connectors\u003C\u002Fstrong> and configure the connector with your endpoint URL (e.g. \u003Ccode>http:\u002F\u002Flocalhost:11434\u002Fv1\u003C\u002Fcode> for Ollama).\u003C\u002Fli>\n\u003Cli>Optionally provide an API key for services that require authentication.\u003C\u002Fli>\n\u003Cli>The plugin registers a provider with the WordPress AI Client and dynamically discovers all available models from your endpoint.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Cp>The plugin also handles practical concerns like extended HTTP timeouts for slow local inference and non-standard port support.\u003C\u002Fp>\n","Connects the WordPress AI Client to Ollama, LM Studio, or any AI endpoint that uses the standard chat completions API format.",90,827,"2026-05-26T16:13:00.000Z","6.9",[18,52,19,20,21],"","https:\u002F\u002Fdownloads.wordpress.org\u002Fplugin\u002Fultimate-ai-connector-compatible-endpoints.2.1.2.zip",{"slug":71,"name":72,"version":73,"author":74,"author_profile":75,"description":76,"short_description":77,"active_installs":78,"downloaded":79,"rating":11,"num_ratings":11,"last_updated":80,"tested_up_to":81,"requires_at_least":82,"requires_php":16,"tags":83,"homepage":86,"download_link":87,"security_score":25,"vuln_count":11,"unpatched_count":11,"last_vuln_date":26,"fetched_at":88},"bix-ai-writer","Bix AI Writer","1.0.3","bilirsoft","https:\u002F\u002Fprofiles.wordpress.org\u002Fbilirsoft\u002F","\u003Cp>\u003Cstrong>Bix AI Writer\u003C\u002Fstrong> adds a simple panel in the Gutenberg editor where you can:\u003Cbr \u002F>\n– Generate full \u003Cstrong>content\u003C\u002Fstrong> for a post\u002Fpage from a prompt.\u003Cbr \u002F>\n– Create a clean \u003Cstrong>summary\u003C\u002Fstrong> from any text you paste.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Bring your own backend:\u003C\u002Fstrong> choose OpenAI, OpenRouter, or a fully free local setup with \u003Cstrong>Ollama\u003C\u002Fstrong> (\u003Ccode>http:\u002F\u002Flocalhost:11434\u003C\u002Fcode>). You keep control of keys and models. No vendor lock-in.\u003C\u002Fp>\n\u003Ch3>Privacy and Data Usage\u003C\u002Fh3>\n\u003Cp>Bix AI Writer connects to third-party AI services to generate content or summaries based on your prompts. Depending on the provider you select in \u003Cstrong>Settings \u003Cspan aria-hidden=\"true\" class=\"wp-exclude-emoji\">→\u003C\u002Fspan> AI Writer\u003C\u002Fstrong>, your text will be sent to:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Ollama (Local, Self-Hosted)\u003C\u002Fstrong> — https:\u002F\u002Follama.com\u002F\u003Cbr \u002F>\nYour data stays on your own server. No external transmission is involved unless your Ollama endpoint is remote.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>OpenAI\u003C\u002Fstrong> — https:\u002F\u002Fopenai.com\u002F — Privacy Policy: https:\u002F\u002Fopenai.com\u002Fprivacy\u003C\u002Fli>\n\u003Cli>\u003Cstrong>OpenRouter\u003C\u002Fstrong> — https:\u002F\u002Fopenrouter.ai\u002F — Privacy Policy: https:\u002F\u002Fopenrouter.ai\u002Fprivacy\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>The plugin itself does \u003Cstrong>not\u003C\u002Fstrong> collect analytics or store personal data; it only inserts the returned text into the editor. API requests are performed from your WordPress server directly to the configured provider. You are responsible for complying with the selected provider’s terms of use and privacy policies.\u003C\u002Fp>\n","AI-assisted content & summary generator for the WordPress block editor. Works with OpenAI, OpenRouter, or local Ollama. No vendor lock-in.",10,290,"2025-10-17T10:40:00.000Z","6.8.5","6.0",[18,84,21,22,85],"content","summary","https:\u002F\u002Fwordpress.org\u002Fplugins\u002Fbix-ai-writer\u002F","https:\u002F\u002Fdownloads.wordpress.org\u002Fplugin\u002Fbix-ai-writer.zip","2026-04-16T10:56:18.058Z",{"slug":90,"name":91,"version":92,"author":93,"author_profile":94,"description":95,"short_description":96,"active_installs":11,"downloaded":97,"rating":11,"num_ratings":11,"last_updated":98,"tested_up_to":14,"requires_at_least":15,"requires_php":16,"tags":99,"homepage":101,"download_link":102,"security_score":25,"vuln_count":11,"unpatched_count":11,"last_vuln_date":26,"fetched_at":27},"ai-provider-for-llamacpp","AI Provider for llama.cpp","0.0.3","AcrossWP","https:\u002F\u002Fprofiles.wordpress.org\u002Facrosswp\u002F","\u003Cp>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.\u003C\u002Fp>\n\u003Cp>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.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Features:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Text generation with any llama.cpp-loaded model\u003C\u002Fli>\n\u003Cli>Automatic model discovery from your llama.cpp server\u003C\u002Fli>\n\u003Cli>Function calling support\u003C\u002Fli>\n\u003Cli>Structured output (JSON mode) support\u003C\u002Fli>\n\u003Cli>Settings page to configure the server URL (default: http:\u002F\u002F127.0.0.1:8080)\u003C\u002Fli>\n\u003Cli>Works without an API key for local instances\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cstrong>Requirements:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>PHP 7.4 or higher\u003C\u002Fli>\n\u003Cli>WordPress AI Client plugin must be installed and activated\u003C\u002Fli>\n\u003Cli>llama.cpp server running locally or on a remote host\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch3>Getting Started\u003C\u002Fh3>\n\u003Ch4>What do I need before using this plugin?\u003C\u002Fh4>\n\u003Cp>You need a running llama.cpp server (local or remote) and the WordPress AI Client plugin installed and activated.\u003C\u002Fp>\n\u003Ch4>How do I install llama.cpp?\u003C\u002Fh4>\n\u003Cp>On macOS run \u003Ccode>brew install llama.cpp\u003C\u002Fcode>. For other platforms, see the official docs at https:\u002F\u002Fllama-cpp.com\u002Fdownload\u002F\u003C\u002Fp>\n\u003Ch4>Where do I get a model?\u003C\u002Fh4>\n\u003Cp>Download a GGUF model from Hugging Face. TinyLlama 1.1B (Q4_K_M, ~636 MB) is a good starting point for testing.\u003C\u002Fp>\n\u003Ch4>How do I start the server?\u003C\u002Fh4>\n\u003Cp>Run \u003Ccode>llama-server --models-dir ~\u002Fmodels\u003C\u002Fcode>. The server starts on \u003Ccode>http:\u002F\u002F127.0.0.1:8080\u003C\u002Fcode> by default.\u003C\u002Fp>\n\u003Ch4>How do I connect WordPress to my llama.cpp server?\u003C\u002Fh4>\n\u003Cp>Go to \u003Cstrong>Settings > llama.cpp\u003C\u002Fstrong> and enter your server URL. Leave it blank to use the default (\u003Ccode>http:\u002F\u002F127.0.0.1:8080\u003C\u002Fcode>).\u003C\u002Fp>\n\u003Cp>For full step-by-step setup instructions, see the \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Facrosswp\u002Fai-provider-for-llamacpp\u002F\" rel=\"nofollow ugc\">Getting Started Guide\u003C\u002Fa>.\u003C\u002Fp>\n","llama.cpp provider for the WordPress AI Client.",410,"2026-04-27T22:32:00.000Z",[18,52,100,19,20],"llamacpp","https:\u002F\u002Fgithub.com\u002FacrossWP\u002Fai-provider-for-llamacpp","https:\u002F\u002Fdownloads.wordpress.org\u002Fplugin\u002Fai-provider-for-llamacpp.0.0.3.zip",{"slug":104,"name":105,"version":106,"author":107,"author_profile":108,"description":109,"short_description":110,"active_installs":11,"downloaded":111,"rating":11,"num_ratings":11,"last_updated":112,"tested_up_to":14,"requires_at_least":15,"requires_php":16,"tags":113,"homepage":115,"download_link":116,"security_score":25,"vuln_count":11,"unpatched_count":11,"last_vuln_date":26,"fetched_at":27},"ai-provider-for-osaurus","AI Provider for Osaurus","0.4.2","Andrei Lupu","https:\u002F\u002Fprofiles.wordpress.org\u002Feuthelup\u002F","\u003Cp>AI Provider for Osaurus wires the \u003Ca href=\"https:\u002F\u002Fosaurus.ai\" rel=\"nofollow ugc\">Osaurus\u003C\u002Fa> local LLM runtime into the WordPress AI Client that ships with WordPress 7.0.\u003C\u002Fp>\n\u003Cp>After activation, any plugin that calls \u003Ccode>wp_ai_client_prompt()\u003C\u002Fcode> (Gutenberg AI features, official provider plugins, third-party code) can route prompts to Osaurus running on the same machine. Prompts never leave the local network.\u003C\u002Fp>\n\u003Ch4>Why use it\u003C\u002Fh4>\n\u003Cul>\n\u003Cli>\u003Cstrong>Privacy first\u003C\u002Fstrong> — prompts stay on your machine, no third-party API receives the data.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Zero cost\u003C\u002Fstrong> — no API keys, no metered requests.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Fast\u003C\u002Fstrong> — MLX-backed inference on Apple Silicon.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>OpenAI-compatible wire format\u003C\u002Fstrong> — streaming, tool calls, JSON output, and chat history work out of the box.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch4>Capabilities\u003C\u002Fh4>\n\u003Cul>\n\u003Cli>Text generation\u003C\u002Fli>\n\u003Cli>Chat history\u003C\u002Fli>\n\u003Cli>Function \u002F tool calls\u003C\u002Fli>\n\u003Cli>Structured JSON output\u003C\u002Fli>\n\u003Cli>Streaming\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Image generation, embeddings, and text-to-speech are not supported because Osaurus does not currently expose those endpoints. Pair this plugin with an image-capable provider such as the official AI Provider for OpenAI when you need images.\u003C\u002Fp>\n\u003Ch4>How it works\u003C\u002Fh4>\n\u003Cp>The plugin registers a provider with \u003Ccode>AiClient::defaultRegistry()\u003C\u002Fcode> and points it at the Osaurus HTTP server. Because Osaurus speaks the OpenAI Chat Completions wire format, the implementation extends the SDK’s OpenAI-compatible base classes — so capabilities like streaming and tool calling flow through automatically.\u003C\u002Fp>\n\u003Ch4>Configuration\u003C\u002Fh4>\n\u003Cp>The plugin resolves the Osaurus base URL in this order (first match wins):\u003C\u002Fp>\n\u003Col>\n\u003Cli>\u003Ccode>OSAURUS_BASE_URL\u003C\u002Fcode> PHP constant defined in \u003Ccode>wp-config.php\u003C\u002Fcode>.\u003C\u002Fli>\n\u003Cli>\u003Ccode>osaurus_ai_connector_base_url\u003C\u002Fcode> option, settable from \u003Cstrong>Settings → Connectors\u003C\u002Fstrong> in wp-admin.\u003C\u002Fli>\n\u003Cli>Default: \u003Ccode>http:\u002F\u002F127.0.0.1:1337\u002Fv1\u003C\u002Fcode> (works out of the box for bare-metal WordPress on the same Mac as Osaurus; Docker setups must override).\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch4>Pick the right URL for your setup\u003C\u002Fh4>\n\u003Cp>WordPress can run in many places — the right base URL depends on how it reaches your Mac, where Osaurus is listening.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Bare-metal WordPress on the same Mac\u003C\u002Fstrong> (MAMP, Laravel Valet, Local, WordPress Studio, native PHP):\u003Cbr \u002F>\n  http:\u002F\u002F127.0.0.1:1337\u002Fv1 \u003Cem>(this is the default — no configuration needed)\u003C\u002Fem>\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Docker-based WordPress on the same Mac\u003C\u002Fstrong> (DDEV, Lando, Docker Desktop):\u003Cbr \u002F>\n  http:\u002F\u002Fhost.docker.internal:1337\u002Fv1\u003C\u002Fli>\n\u003Cli>\u003Cstrong>WordPress on a different machine on your LAN, Osaurus on your Mac:\u003C\u002Fstrong>\u003Cbr \u002F>\n  http:\u002F\u002F:1337\u002Fv1 — make sure Osaurus is bound to a non-loopback interface and your firewall allows port 1337.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Remote Osaurus you operate\u003C\u002Fstrong> (e.g. a Mac mini on Tailscale, a colocated server):\u003Cbr \u002F>\n  http:\u002F\u002F:1337\u002Fv1 or \u003Ccode>https:\u002F\u002F...\u003C\u002Fcode> if you front it with TLS.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>The fastest way to confirm a URL works from the WordPress host: run \u003Ccode>curl \u003Cbase-url>\u002Fmodels\u003C\u002Fcode> from the same shell environment WordPress runs in. A JSON list back means the connector will work.\u003C\u002Fp>\n\u003Cp>Osaurus project home: \u003Ca href=\"https:\u002F\u002Fosaurus.ai\" rel=\"nofollow ugc\">osaurus.ai\u003C\u002Fa>\u003Cbr \u002F>\nOsaurus documentation: \u003Ca href=\"https:\u002F\u002Fdocs.osaurus.ai\" rel=\"nofollow ugc\">docs.osaurus.ai\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch3>External services\u003C\u002Fh3>\n\u003Cp>This plugin connects to an Osaurus HTTP server — a local LLM runtime that you install and run yourself on the same machine as WordPress (or on a host you operate). The plugin is useless without it: every text-generation request from the WordPress AI Client is routed to this server.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Service: Osaurus (self-hosted, local LLM runtime)\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>Osaurus is an open-source Apple Silicon LLM runtime that exposes an OpenAI-compatible HTTP API. It is installed and operated by the same person running WordPress — there is no third-party service operator. This plugin forwards prompts to it so WordPress can perform text generation, chat, tool calls, and structured JSON output without sending data to any cloud provider.\u003C\u002Fp>\n\u003Cp>Project homepage and download (\u003Ccode>osaurus.ai\u003C\u002Fcode>): https:\u002F\u002Fosaurus.ai\u003Cbr \u002F>\nProject documentation (\u003Ccode>docs.osaurus.ai\u003C\u002Fcode>): https:\u002F\u002Fdocs.osaurus.ai\u003C\u002Fp>\n\u003Cp>These URLs are referenced in the plugin’s admin UI as documentation\u002Fdownload links only; the plugin does \u003Cstrong>not\u003C\u002Fstrong> make HTTP requests to \u003Ccode>osaurus.ai\u003C\u002Fcode> or \u003Ccode>docs.osaurus.ai\u003C\u002Fcode>.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Hosts the plugin actually contacts at runtime\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>The plugin sends HTTP requests only to the host and port resolved (in order) from:\u003C\u002Fp>\n\u003Col>\n\u003Cli>The \u003Ccode>OSAURUS_BASE_URL\u003C\u002Fcode> PHP constant (if defined in \u003Ccode>wp-config.php\u003C\u002Fcode>).\u003C\u002Fli>\n\u003Cli>The \u003Ccode>osaurus_ai_connector_base_url\u003C\u002Fcode> WordPress option (set from \u003Cstrong>Settings → Connectors\u003C\u002Fstrong>).\u003C\u002Fli>\n\u003Cli>The built-in default \u003Ccode>http:\u002F\u002F127.0.0.1:1337\u002Fv1\u003C\u002Fcode>.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Cp>Two host targets are referenced by name in the plugin code as quick-pick presets and as the default:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Ccode>127.0.0.1:1337\u003C\u002Fcode> — loopback on the same machine as WordPress (default). Used by bare-metal installs (Studio, MAMP, Valet, Local, native PHP).\u003C\u002Fli>\n\u003Cli>\u003Ccode>host.docker.internal:1337\u003C\u002Fcode> — Docker’s hostname for the host machine. Used by Docker-based WordPress (DDEV, Lando, Docker Desktop). Resolved by Docker, not by a public DNS server.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>If the user configures a different URL, the plugin only contacts that URL. The plugin never contacts any third-party endpoint operated by the plugin author, the Osaurus project, or any other party.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>What data is sent and when\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>When any plugin (including WordPress core) calls \u003Ccode>wp_ai_client_prompt()\u003C\u002Fcode> and routes through Osaurus, the plugin sends a \u003Ccode>POST\u003C\u002Fcode> to \u003Ccode>{base_url}\u002Fchat\u002Fcompletions\u003C\u002Fcode> with: the prompt text, the conversation history you supplied, the model ID, sampling parameters (temperature, top_p, max tokens, etc.), and any tool \u002F function declarations or JSON schemas you supplied.\u003C\u002Fli>\n\u003Cli>When an admin opens \u003Cstrong>Settings → Connectors\u003C\u002Fstrong>, the plugin sends a \u003Ccode>GET {base_url}\u002Fmodels\u003C\u002Fcode> to populate the model picker and a \u003Ccode>GET {base_root}\u002Fhealth\u003C\u002Fcode> to display a connection status indicator. Both run only for users with \u003Ccode>manage_options\u003C\u002Fcode> and are triggered by user actions in wp-admin.\u003C\u002Fli>\n\u003Cli>No telemetry, analytics, or background requests are sent.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cstrong>Terms of service and privacy policy\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>Osaurus is open-source software you self-host. There is no third-party service operator collecting your data. See the project pages linked above for the project’s terms and source. If you point the plugin at a remote Osaurus server operated by someone else (for example, a Mac mini you reach over Tailscale), the operator of that server is the data recipient and is governed by whatever terms you have with them.\u003C\u002Fp>\n","Registers Osaurus, a local Apple Silicon LLM runtime, as a provider for the WordPress AI Client.",116,"2026-06-02T01:35:00.000Z",[18,52,19,20,114],"osaurus","https:\u002F\u002Fgithub.com\u002Fandreilupu\u002Fai-provider-for-osaurus","https:\u002F\u002Fdownloads.wordpress.org\u002Fplugin\u002Fai-provider-for-osaurus.0.4.2.zip",{"error":118,"url":119,"statusCode":120,"statusMessage":121,"message":121},true,"http:\u002F\u002Flocalhost\u002Fapi\u002Fplugins\u002Fai-provider-for-openai-compatible-servers\u002Fbundle",404,"no bundle for this plugin yet",{"slug":4,"current_version":6,"total_versions":123,"versions":124},1,[125],{"version":6,"download_url":24,"svn_tag_url":126,"released_at":26,"has_diff":127,"diff_files_changed":128,"diff_lines":26,"trac_diff_url":26,"vulnerabilities":129,"is_current":118},"https:\u002F\u002Fplugins.svn.wordpress.org\u002Fai-provider-for-openai-compatible-servers\u002Ftags\u002F1.0.0\u002F",false,[],[]]