=== AI Provider for NVIDIA ===
Contributors: deepakbhojwani
Tags: ai, nvidia, llama, ai-provider, connector
Requires at least: 6.9
Tested up to: 7.0.2
Stable tag: 1.0.0
Requires PHP: 7.4
License: GPL-2.0-or-later
License URI: https://www.gnu.org/licenses/gpl-2.0.html

AI Provider for NVIDIA NIM for the PHP AI Client SDK. Free API keys unlock Llama, Nemotron, Mistral, Qwen and FLUX image models.

== Description ==

This plugin adds NVIDIA NIM (NVIDIA Inference Microservices) as a provider for the PHP AI Client SDK. It enables WordPress sites and plugins built on the AI Client to use NVIDIA-hosted models for text and image generation.

NVIDIA NIM offers free API credits on signup, and a single API key works across dozens of hosted chat models (Llama, Nemotron, Mistral, Qwen, GPT-OSS, and more) as well as FLUX image generation models.

**Features:**

* Text generation with NVIDIA-hosted chat models (Llama, Nemotron, Mistral, Qwen, GPT-OSS, and more)
* Image generation with NVIDIA-hosted FLUX models (FLUX.1 [dev], FLUX.1 [schnell], FLUX.2 Klein)
* Multimodal image input on vision-capable models
* JSON / structured output support
* Function calling (tool) support
* Reasoning ("thinking") model support
* Automatic provider registration with the PHP AI Client

The provider exposes a curated set of NVIDIA NIM models that are verified to be available for inference. The NVIDIA model catalog lists many models that are not actually provisioned for a given account; surfacing only verified models ensures every model you can choose actually works.

**Requirements:**

* PHP 7.4 or higher
* For WordPress 6.9, the [wordpress/php-ai-client](https://github.com/WordPress/php-ai-client) package must be installed
* For WordPress 7.0 and above, no additional changes are required
* A free NVIDIA NIM API key

This plugin is a provider add-on: on its own it does nothing visible. It only takes effect once the PHP AI Client is available and another plugin (or your code) makes an AI request using the `nvidia` provider.

== Installation ==

1. Upload the plugin files to `/wp-content/plugins/ai-provider-for-nvidia/`, or install through the WordPress plugins screen.
2. Activate the plugin through the 'Plugins' menu in WordPress.
3. Ensure the PHP AI Client is available (bundled with WordPress 7.0, or via the WordPress AI plugin / the wordpress/php-ai-client package on WordPress 6.9).
4. Configure your NVIDIA API key via the `NVIDIA_API_KEY` environment variable or PHP constant.

== Frequently Asked Questions ==

= How do I get a NVIDIA API key? =

Visit [build.nvidia.com](https://build.nvidia.com), create a free account, and generate an API key. NVIDIA keys start with `nvapi-` and include free inference credits.

= Does this plugin work without the PHP AI Client? =

No. This plugin requires the PHP AI Client. It provides the NVIDIA-specific implementation that the PHP AI Client uses, and registers itself automatically when the client is available.

= Which models are supported? =

A curated set of NVIDIA NIM chat models (Llama, Nemotron, Mistral, Qwen, GPT-OSS, Gemma, and more), including vision-capable models that also accept image input, plus NVIDIA-hosted FLUX image generation models. The provider verifies each listed chat model against NVIDIA's live catalog so unavailable models are not shown.

= Does this plugin generate images? =

Yes. It supports NVIDIA-hosted FLUX image generation models (FLUX.1 [dev], FLUX.1 [schnell], and FLUX.2 Klein) through NVIDIA's GenAI endpoint. Vision-capable chat models can additionally accept images as input.

= Is this plugin affiliated with NVIDIA? =

No. NVIDIA is a trademark of NVIDIA Corporation. This plugin is an independent integration and is not affiliated with, endorsed by, or sponsored by NVIDIA Corporation.

== External services ==

This plugin connects to the NVIDIA NIM API to provide AI text and image generation. It is required for the plugin's core functionality and only runs when you (or a plugin using the PHP AI Client) trigger an AI request.

- What it does: sends your prompt text, model selection, and generation parameters to NVIDIA's inference API and receives the generated text or image.
- When: only when an AI request is made via the PHP AI Client using the `nvidia` provider. No data is sent on page load or in the background.
- Data sent: the prompt/messages you provide, the chosen model ID, generation parameters, and your NVIDIA API key (as an Authorization header) for authentication.
- Services: NVIDIA NIM, which uses two endpoints:
  - `integrate.api.nvidia.com` — text generation (chat models) and model discovery.
  - `ai.api.nvidia.com` — image generation (FLUX models).
  - Terms of Service: https://www.nvidia.com/en-us/about-nvidia/terms-of-service/
  - Privacy Policy: https://www.nvidia.com/en-us/about-nvidia/privacy-policy/

NVIDIA is a trademark of NVIDIA Corporation. This plugin is an independent integration and is not affiliated with, endorsed by, or sponsored by NVIDIA Corporation.

== Changelog ==

= 1.0.0 =

* Initial release.
* Text generation with NVIDIA NIM chat models via the OpenAI-compatible Chat Completions API.
* Image generation with NVIDIA-hosted FLUX models via the NVIDIA GenAI API.
* Curated, verified model list — only models that are actually provisioned for inference are surfaced (the NVIDIA catalog also lists many models that are not callable per account).
* Image input on vision models, JSON output, function calling, and reasoning support.
