WP Manifestindependent plugin directory
manifest / ai / ai-provider-for-tencent-cloud-token-hub

AI Provider for Tencent Cloud Token Hub

WordPress AI Provider for Tencent Cloud Token Hub

by Bestony · github.com/bestony/ai-provider-for-tencent-cloud-token-hub · website

★ 0stars
0forks

Install

No release zip yet. The repository archive installs, but the folder name will carry the branch suffix and updates will not flow:

wp plugin install https://github.com/bestony/ai-provider-for-tencent-cloud-token-hub/archive/refs/heads/main.zip
Readme github.com/bestony/ai-provider-for-tencent-cloud-token-hub ↗

Tencent Cloud Token Hub as an OpenAI-compatible provider for the WordPress AI Client.

What it does

  • Discovers models from GET /models and keeps only entries whose status is online.
  • Sends text, chat history, tool calls, streaming requests, and structured output to POST /chat/completions.
  • Declares image input only for the maintained list of confirmed Token Hub vision models.
  • Excludes image, video, 3D generation, embedding, speech, and other non-chat models because this plugin implements Chat Completions only.
  • Uses the AI Client credential registry for the Connector key. The key never passes through this plugin's configuration code.

The provider ID is tencentcloud_tokenhub.

Requirements

  • WordPress 7.0 or newer with the PHP AI Client SDK.
  • PHP 7.4 or newer.
  • A Tencent Cloud Token Hub API key.

Install

Copy this directory to wp-content/plugins/ai-provider-for-tencentcloud-tokenhub/, activate it, then open Settings → Connectors. Open Tencent Cloud Token Hub and enter the API key.

Configuration

Credentials can be supplied through the AI Client Connector (connectors_ai_tencentcloud_tokenhub_api_key) or through the SDK-supported TENCENTCLOUD_TOKENHUB_API_KEY environment variable or PHP constant. This provider checks the AI Client registry and does not read the option directly.

Setting Default Purpose
TENCENTCLOUD_TOKENHUB_DEFAULT_MODEL hy4-preview Model placed first in text and vision model preferences
TENCENTCLOUD_TOKENHUB_REQUEST_TIMEOUT 120 Request timeout in seconds
TENCENTCLOUD_TOKENHUB_CONNECT_TIMEOUT 10 Connection timeout in seconds
TENCENTCLOUD_TOKENHUB_STRUCTURED_OUTPUT json_schema json_schema, json_object, or none

The API base URL is fixed at https://api.lkeap.cloud.tencent.com/plan/v3. It is used for both model discovery and generation so the API key is not sent to a configurable host.

Token Hub accepts at most four messages in one Chat Completions request. The provider rejects a longer prompt before making a network request. Split long conversations before calling the AI Client.

The /models response does not include capability metadata. Unknown online models are exposed as text-only chat models. The catalog contains only confirmed vision rules; update it when Tencent documents a new vision model. Image, video, and 3D generation models remain visible to the upstream service but are excluded from this text-only provider.

Model preferences

The provider adds hy4-preview to the front of the standard preference lists when a credential is configured. Other provider entries remain in their original order:

add_filter('wpai_preferred_text_models', function ($models) {
    array_unshift($models, ['tencentcloud_tokenhub', 'hy4-preview']);
    return $models;
});

Endpoints and data

The plugin calls:

  • GET https://api.lkeap.cloud.tencent.com/plan/v3/models to discover online models.
  • POST https://api.lkeap.cloud.tencent.com/plan/v3/chat/completions to generate text.

Both requests use Authorization: Bearer <API key>. Prompts, conversation history, tool definitions, schemas, and attached image data supplied by the calling plugin are sent to Token Hub when a generation is requested.

Development

php scripts/selfcheck.php
php scripts/selfcheck.php --sdk=/path/to/wordpress-php-ai-client/src

The self-check never needs a real API key and never calls Tencent Cloud.

License

GPL-2.0-or-later. See LICENSE.