WordPress.org

Taiwan 正體中文

取得 WordPress
WordPress.org

Plugin Directory

AI

Image
Image

AI

由 WordPress.org 開發
下載
即時預覽

外掛說明

The AI plugin adds opt-in AI tools to the WordPress admin and Block Editor. It’s built by the WordPress Core AI team for site owners, content creators, and developers who want to try AI in WordPress early and help shape what comes next.

Features and Experiments

Nothing is turned on when you activate the plugin. You choose what to enable in Settings -> AI, and each item there is one of two kinds:

  • Feature – Stable and supported. Changes avoid breaking existing behavior. Image Generation and Editing is currently the only Feature.
  • Experiment – Early-stage and actively being developed. Experiments may change significantly or be removed in any release, so we recommend trying them on a staging site before using them on a live site.

Once enabled, most Experiments only send a request to your AI provider when you click a button, but some run on their own. Comment Moderation, for example, checks every new comment, including comments from site visitors. In Settings -> AI, Experiments that run on their own are marked “Runs automatically”, and the info icon next to each item’s name shows when it runs, what it sends to your AI provider, and how many requests it makes.

Read the Feature and Experiment Lifecycle to learn how Experiments are evaluated and how they can become Features.

Before You Start

  • This plugin is not part of WordPress core. It does nothing until you install and activate it, and each Feature or Experiment stays off until you enable it.
  • You need an AI Connector plugin and an account with an AI provider. The AI plugin doesn’t include provider credentials. Install a connector plugin such as Anthropic, Google, or OpenAI (more connectors), then add your API key in Settings -> Connectors.
  • Your AI provider bills you for usage. The plugin is free; each AI request is billed by your provider under your account.
  • The Block Editor is required. The Classic Editor and other editing experiences are not supported.

New to WordPress AI? Take the AI-Powered WordPress course to learn how to configure the plugin, use its features, and understand the building blocks behind them. The Getting Started guide covers setup and troubleshooting.

Feature

  • Image Generation and Editing – Create and edit images from post content in the editor, and from the Media Library.

Experiments

Experiments marked “runs automatically” do their work without you clicking anything once they’re enabled. Most of them send requests to your AI provider when they run.

  • Abilities Explorer – Browse and test registered AI abilities from a dedicated admin screen.
  • AI Request Logging – Logs AI requests, including short previews of prompts and responses, for observability and debugging.
  • Alt Text Generation – Generate descriptive alt text for images to improve accessibility.
  • Block Container Label Generation – Generates short List View labels for Group, Columns, and other container blocks.
  • Comment Moderation (runs automatically) – Checks each new comment for toxicity and sentiment, gives it a value score, and can hold comments for moderation. Includes comments from site visitors by default.
  • Connector Approval – Require administrator approval before other plugins or themes can use the AI connectors configured on this site.
  • 內容分類 – 建議相關的標籤與分類，以利整理內容。
  • Content Resizing – Shorten, expand, or rephrase selected block content.
  • Content Summarization – Summarizes long-form content into digestible overviews.
  • Content Translation – Translates paragraph and heading blocks, and optionally the post title, into a selected language from the post editor.
  • Content Verification (runs automatically) – Detects C2PA Content Credentials in uploaded images and stores them alongside the attachment.
  • Custom Abilities – Makes the plugin’s general-purpose WordPress Abilities available through the Abilities API and MCP.
  • Editorial Notes – Reviews post content block-by-block and adds Notes with suggestions for Accessibility, Readability, Grammar, and SEO.
  • Editorial Updates – Applies pending editorial Notes to your content and saves the post, without a preview.
  • Excerpt Generation – Generates excerpt suggestions for your posts.
  • Key Encryption – Encrypts AI provider API keys stored in your database. Disabling the experiment or deactivating the plugin restores plaintext keys.
  • Markdown Feeds – Publishes a Markdown feed and Markdown versions of your public posts and pages for AI agents and other machine readers.
  • MCP Access – Controls which abilities AI agents can use over the Model Context Protocol (MCP). Installs and activates the MCP Adapter plugin when enabled.
  • Meta Description Generation – Generates meta description suggestions and integrates them with several SEO plugins.
  • Revision Change Summary – Summarizes what changed between two revisions of a post in plain language, on the visual and classic revisions screens.
  • Slug Generation – Suggests SEO-friendly permalink slugs from the permalink popover or the pre-publish panel.
  • Suggest Reply – Adds a “Suggest Reply” action to the Comments screen and Activity widget to draft replies to comments.
  • Text to Speech – Generates an audio version of a post that visitors can listen to instead of reading.
  • Title Generation – Generates title suggestions for your posts.
  • Type-ahead Text (runs automatically) – Suggests text as you type in the Block Editor.
  • WebMCP – Lets an agent browser work in the Block Editor through WebMCP, with every change visible on the page as it happens.

Coming Soon

我們正積極開發新功能，以強化你的 WordPress 工作流程：

  • AI 試驗場 – 試用不同的 AI 模型及供應商。
  • 內容助理 – Gutenberg 中由 AI 輔助的寫作及編輯功能。
  • 網站代理 – 使用自然語言管理 WordPress。
  • 工作流程自動化 – 由 AI 驅動的工作自動化。

Roadmap

你可以在 WordPress AI GitHub 專案看板的篩選檢視中，查看外掛目前的開發藍圖。

開發者專區

AI 外掛的設計可供研究、擴充，並在其基礎上進一步開發。無論你是外掛開發者、代理商或主機服務供應商，都可以進行以下工作：

Under the Hood:

These building blocks power the Features and Experiments above. They aren’t settings you turn on or off.

  • AI Building Blocks – Combines the PHP AI Client and Abilities API into one plugin, as a reference implementation for developers.
  • Dashboard Widgets – AI Status and AI Capabilities widgets, plus a framework for registering new ones.
  • Experiment Framework – The opt-in system that registers each Feature and Experiment and its toggle.
  • Guidelines – Lets abilities respect site-wide editorial standards.
  • Multi-Provider Support – Works with AI Connector plugins for providers such as OpenAI, Google, and Anthropic.

擴充外掛：

  • 建立自訂實驗功能 – 使用 Abstract_Feature 基底類別，建立自己的 AI 功能。
  • 預先設定供應商 – 主機服務商及代理商可設定 AI Connector 外掛，讓使用者不必自行準備 API 金鑰。
  • 註冊自訂能力 – 串接 Abilities API 以加入新的 AI 功能。
  • 覆寫預設行為 – 使用篩選器自訂提示、回應及使用者介面元素。
  • 完整的 hook 支援 – 整個程式碼庫均提供可供自訂使用的篩選器及動作。

即將推出的開發者工具：

  • AI 試驗場 – 試用不同的 AI 模型與提示詞。
  • MCP (Model Context Protocol) – 在 WordPress 工作流程中整合並測試 Model Context Protocol 的功能。
  • 擴充供應商 – 支援以額外或替代的 AI 供應商進行實驗。

開始使用：

  1. 閱讀參與貢獻指南，瞭解如何設定開發環境
  2. 前往 WordPress Slack 的 #core-ai 參與討論
  3. 瀏覽 GitHub 儲存庫，瞭解實驗功能的建置方式
  4. 參與討論，交流外掛後續該如何改進。

歡迎參與貢獻！無論你想建立新的實驗、改善現有功能，或協助撰寫文件，都可前往我們的 GitHub 儲存庫參與。

螢幕擷圖

功能：圖片產生與編輯。文章編輯器側邊欄顯示 [產生精選圖片] 按鈕，以及已產生的精選圖片預覽，並包含替代文字、標題及內容說明。
功能：圖片產生與編輯。文章編輯器側邊欄顯示 [產生精選圖片] 按鈕，以及已產生的精選圖片預覽，並包含替代文字、標題及內容說明。
編輯器實驗功能：替代文字產生。圖片區塊設定顯示 [產生替代文字] 按鈕及產生的替代文字。
編輯器實驗功能：替代文字產生。圖片區塊設定顯示 [產生替代文字] 按鈕及產生的替代文字。
編輯器實驗功能：替代文字產生。在媒體庫中批次產生替代文字。
編輯器實驗功能：替代文字產生。在媒體庫中批次產生替代文字。
編輯器實驗功能：留言審核。留言管理畫面顯示 AI 輔助的留言審核功能，包括以顏色區分的標章所呈現的毒性評分及留言情緒。
編輯器實驗功能：留言審核。留言管理畫面顯示 AI 輔助的留言審核功能，包括以顏色區分的標章所呈現的毒性評分及留言情緒。
編輯器實驗功能：留言審核。[活動] 小工具的 [近期留言] 區段顯示情緒及毒性分數。
編輯器實驗功能：留言審核。[活動] 小工具的 [近期留言] 區段顯示情緒及毒性分數。
編輯器實驗功能：內容分類。AI 會分析內容，提供文章標籤及分類的建議。
編輯器實驗功能：內容分類。AI 會分析內容，提供文章標籤及分類的建議。
編輯器實驗功能：內容摘要。文章編輯器側邊欄顯示 [產生 AI 摘要] 按鈕，以及 [內容摘要] 區塊中產生的內容摘要。
編輯器實驗功能：內容摘要。文章編輯器側邊欄顯示 [產生 AI 摘要] 按鈕，以及 [內容摘要] 區塊中產生的內容摘要。
編輯器實驗功能：內容翻譯。可在文章編輯器中，將段落與標題區塊翻譯成選取的語言，也可選擇一併翻譯文章標題。
編輯器實驗功能：內容翻譯。可在文章編輯器中，將段落與標題區塊翻譯成選取的語言，也可選擇一併翻譯文章標題。
編輯器實驗功能：內容摘要產生。文章編輯器側邊欄顯示 [產生內容摘要] 按鈕及產生的內容摘要。
編輯器實驗功能：內容摘要產生。文章編輯器側邊欄顯示 [產生內容摘要] 按鈕及產生的內容摘要。
編輯器實驗功能：中繼描述產生。產生中繼描述建議，並將這些建議整合至各種 SEO 外掛。
編輯器實驗功能：中繼描述產生。產生中繼描述建議，並將這些建議整合至各種 SEO 外掛。
代稱產生：根據文章標題或內容，在永久連結彈出面板及發佈前面板中提供有利於 SEO 的永久連結代稱建議。
代稱產生：根據文章標題或內容，在永久連結彈出面板及發佈前面板中提供有利於 SEO 的永久連結代稱建議。
編輯器實驗功能：標題產生。文章編輯器在文章標題欄位上方顯示 [產生] 按鈕，並在互動視窗中顯示標題建議。
編輯器實驗功能：標題產生。文章編輯器在文章標題欄位上方顯示 [產生] 按鈕，並在互動視窗中顯示標題建議。
編輯器實驗功能：自動完成文字。在區塊編輯器中撰寫段落時顯示淡色文字建議。
編輯器實驗功能：自動完成文字。在區塊編輯器中撰寫段落時顯示淡色文字建議。
控制台小工具。AI 能力小工具會顯示能力總管摘要、已連線的 AI 供應商及模型能力。
控制台小工具。AI 能力小工具會顯示能力總管摘要、已連線的 AI 供應商及模型能力。
控制台小工具。AI 狀態小工具顯示三步驟設定流程。
控制台小工具。AI 狀態小工具顯示三步驟設定流程。
控制台小工具。AI 狀態小工具顯示已連線的 AI 供應商，以及已啟用的功能及實驗功能。
控制台小工具。AI 狀態小工具顯示已連線的 AI 供應商，以及已啟用的功能及實驗功能。
管理介面實驗功能：能力總管。能力總管管理畫面列出可用的 AI 能力，並提供篩選功能、供應商資訊及測試操作。
管理介面實驗功能：能力總管。能力總管管理畫面列出可用的 AI 能力，並提供篩選功能、供應商資訊及測試操作。
管理介面實驗功能：能力總管。能力總管的詳細資料檢視畫面會顯示 AI 能力的說明、供應商、輸入結構描述、輸出結構描述及原始資料。
管理介面實驗功能：能力總管。能力總管的詳細資料檢視畫面會顯示 AI 能力的說明、供應商、輸入結構描述、輸出結構描述及原始資料。
管理介面實驗功能：能力總管。能力總管的測試能力畫面會顯示 AI 能力的 JSON 輸入資料、驗證結果及輸入結構描述參考。
管理介面實驗功能：能力總管。能力總管的測試能力畫面會顯示 AI 能力的 JSON 輸入資料、驗證結果及輸入結構描述參考。
管理介面實驗功能：AI 請求記錄。記錄 AI 請求，以供可觀測性及偵錯使用。可在 [工具] 中檢視詳細記錄。
管理介面實驗功能：AI 請求記錄。記錄 AI 請求，以供可觀測性及偵錯使用。可在 [工具] 中檢視詳細記錄。
管理介面實驗功能：AI Connector 核准。外掛或佈景主題必須先取得網站管理員明確核准，才能使用這個網站已設定的 AI Connector。
管理介面實驗功能：AI Connector 核准。外掛或佈景主題必須先取得網站管理員明確核准，才能使用這個網站已設定的 AI Connector。
管理介面實驗功能：金鑰加密。使用內附的 libsodium 加密功能，將靜態儲存的 AI 供應商 API 金鑰加密。讀取時會自動解密，寫入時則會重新加密。停用此實驗功能或停用外掛後，金鑰會還原為純文字。
管理介面實驗功能：金鑰加密。使用內附的 libsodium 加密功能，將靜態儲存的 AI 供應商 API 金鑰加密。讀取時會自動解密，寫入時則會重新加密。停用此實驗功能或停用外掛後，金鑰會還原為純文字。
管理介面實驗功能：建議回覆。在 [留言] 畫面及 [活動] 小工具加入「建議回覆」操作，讓審核人員能快速產生留言回覆建議。
管理介面實驗功能：建議回覆。在 [留言] 畫面及 [活動] 小工具加入「建議回覆」操作，讓審核人員能快速產生留言回覆建議。
AI 設定。AI 設定畫面顯示可啟用特定實驗的切換開關。
AI 設定。AI 設定畫面顯示可啟用特定實驗的切換開關。
開發人員工具：匯出及匯入設定選項。
開發人員工具：匯出及匯入設定選項。

安裝方式

  1. Install the plugin through the Plugins -> Add Plugin screen, or upload the plugin files to the /wp-content/plugins/ai directory. We recommend trying the plugin on a staging site first.
  2. Activate the plugin through the Plugins screen in WordPress.
  3. Install and activate at least one AI Connector plugin, then go to Settings -> Connectors and add your API key.
  4. Go to Settings -> AI and enable the Features or Experiments you want to use. Select the info icon next to each Experiment’s name to see when it runs and what it sends to your AI provider.
  5. Try it out. For Title Generation, edit a post and click into the title field. Click the Generate button above the field, then choose a suggestion and click Select to insert it.

常見問題集

這個外掛有什麼用途？

This plugin brings opt-in AI tools for writing, editing, media, and site administration into WordPress. It’s where the WordPress Core AI team builds and tests AI features with the community, and it’s a reference implementation for developers who want to build their own AI features.

Is this plugin part of WordPress core?

No. The AI plugin is a separate plugin that you choose to install and activate. WordPress core includes the underlying AI Client and the Settings -> Connectors screen, but it doesn’t send anything to an AI provider on its own. AI Connector plugins (for Anthropic, Google, OpenAI, and others) are also separate plugins, and you can deactivate and delete them at any time.

What’s the difference between a Feature and an Experiment?

Features are stable and supported, and changes to them avoid breaking existing behavior. Experiments are early-stage: they may change significantly or be removed in any release. Both are off by default. See the Feature and Experiment Lifecycle for details.

在正式網站使用這個外掛安全嗎？

Image Generation and Editing is a stable Feature that only runs when you ask for it in the editor or Media Library.

Experiments may change or be removed in any release, so we recommend trying them on a staging site before enabling them on a live site. Once enabled, most Experiments only run when you click a button. A few run automatically:

  • Comment Moderation sends every new comment, including comments from site visitors, to your AI provider and can hold comments for moderation.
  • Type-ahead Text requests suggestions while you type in the Block Editor.
  • Content Verification checks each new image upload for Content Credentials. It doesn’t send anything to your AI provider.

Some Experiments change your content without a preview after you click their button. For example, Editorial Updates rewrites blocks and saves the post, and Content Translation replaces block text in the editor. You can review or undo saved changes with revisions.

In Settings -> AI, the info icon next to each Experiment’s name shows when it runs and what it sends to your AI provider.

支援哪些 AI 供應商？

The plugin can work with provider connector plugins from Anthropic (Claude), Google (Gemini), OpenAI, and others. Install and activate the relevant connector plugin, then configure it in Settings -> Connectors. The AI plugin also lists a few reviewed connector plugins there, such as Mistral and Vercel AI Gateway, so they can be installed from that screen.

需要 API 金鑰才能使用這些功能嗎？

是，目前必須為已設定的 AI Connector 外掛提供自己的 API 金鑰，例如 OpenAI、Google AI 或 Anthropic。

費用是多少？

外掛本身免費，但仍需支付所選 AI 供應商的 API 用量費用。費用因供應商及用量而異。多數供應商會提供免費試用額度供你開始使用。也可使用部分在本機執行、採用開放原始碼且免費的供應商（例如 Ollama）。

Experiments that run automatically can use requests without anyone clicking a button. Comment Moderation, for example, makes one request for each new comment, including comments from site visitors and spam that your spam filter hasn’t already caught. Check your provider’s usage dashboard after enabling an Experiment, and turn on AI Request Logging to see each request the plugin makes.

沒有程式設計知識也能使用嗎？

Yes. The plugin is designed for content creators and site administrators. Once your AI Connectors are configured, you can enable Features and Experiments in Settings -> AI and use them directly from the post editor. As with any Experiment, we recommend trying them on a staging site first.

這個外掛支援 Classic Editor 嗎？

不支援。

AI 外掛目前僅支援區塊編輯器（亦稱 Gutenberg）。此外掛的設計以現代化的編輯器 API、區塊式內容工作流程，以及 WordPress 核心（包括 Gutenberg）持續發展中的編輯功能為基礎。

自 WordPress 5.0 於 2018 年推出以來，區塊編輯器一直是 WordPress 的預設編輯體驗，至今仍是編輯器積極開發的主要焦點。將開發資源集中在區塊編輯器，能讓此專案更快推出新功能、實驗及整合，同時避免在多種編輯體驗中維護平行實作的複雜性。

雖然 Classic Editor 外掛仍有相當龐大的安裝量，Core AI 團隊仍決定優先在區塊編輯器生態系中推動創新及實驗。目前沒有新增 Classic Editor 官方支援的計畫。

Something went wrong. How do I turn the plugin off?

Go to Plugins and deactivate AI. If you can’t reach the admin, deactivate it with WP-CLI (wp plugin deactivate ai) or rename the wp-content/plugins/ai folder using FTP or your host’s file manager. To turn off a single Experiment instead, disable it in Settings -> AI. The Getting Started guide has more troubleshooting steps.

哪裡可以取得協助或回報問題？

使用者評論

Image

So many issues!!!

2026 年 10 月 12 日 7 則留言
1 – It was shipped broken as the WP-CLI has issues with truncating files names and paths to 100 chars. The AI folder names are huge and break this so truncated php files in the folder. 2 – Doesnt take the Anthropic APi keys (yes i have them, yes they are paid for , yes they work other places). NO decent error message as to what is wrong. 3 – ON failing to accept keys, cancelling breaks Wordpress with error messages and locks up the site so you have to FTP into files and disable the plugin by hand then clean up the mess. This is not even close to being ready for use in production servers.
Image

Quite useful, but…

2026 年 9 月 13 日 1 則留言
It is quite useful and performs well, but why did you change the language from “Basic summary” to TL;DR as the title for content summarization? Not everyone who visits our site is a teenager on social media, 25 hours per day.
Image

Doesn’t work with Classic Editor

2026 年 6 月 3 日 1 則留言
I love the idea in theory, and connecting my OpenAI account was easy enough. I was hoping to be able to use the plugin to help with content writing, rather than copying and pasting from ChatGPT. As I use Elementor Page Builder I have the Classic Editor plugin installed – unfortunately this plugin doesn't work with the Classic Editor, so I can't see any of the AI features when writing Posts. Hopefully there will be improvements in the future 🙂
Image

needed

2026 年 5 月 24 日 1 則留言
needed for so long
閱讀全部 10 則使用者評論

參與者及開發者

以下人員參與了開源軟體〈AI〉的開發相關工作。

參與者

〈AI〉外掛目前已有 19 個本地化語言版本。 感謝全部譯者為這個外掛做出的貢獻。

將〈AI〉外掛本地化為台灣繁體中文版

對開發相關資訊感興趣？

任何人均可瀏覽程式碼、查看 SVN 存放庫，或透過 RSS 訂閱開發記錄。

變更記錄

1.5.0 – 2026-10-29

新增

  • New Experiment: Content Verification, that adds read-only detection of C2PA Content Credentials in uploaded JPEG/PNG/WebP images (#459).
  • New Experiment: Text to Speech, that allows you to generate audio for a specific post and then render audio controls on the frontend (#888).
  • New Experiment: MCP Access, that installs and activates the MCP Adapter plugin from WordPress.org when enabled, and adds a Tools → MCP Access screen to control which abilities are exposed to AI agents over MCP (#992).
  • New Experiment: WebMCP, that lets an agent browser work in the block editor through document.modelContext, with tools that set the title, insert and edit blocks, save and publish, with every change visible on the page as it happens (#1081).
  • New Experiment: Revision Change Summary, that summarizes what changed between two revisions of a post on the visual and classic revisions screens (#1159).
  • New Experiment: Block Container Label Generation, that generates concise block container labels in List View and the block Rename modal (#1175, (#1218).
  • A post’s Markdown document (?output_format=markdown) has approved comments appended (#1121).
  • New background sync processing to handle generating embeddings for posts and terms (#1135).
  • “Generate title” button in the post Rename modal (#1140).
  • Learn WordPress AI course links to the plugin readmes and the AI settings header for easier access to course resources (#1154).
  • A neutral badge icon to thumbnails of images with C2PA data detected in the Media Library grid view (#1164).
  • An ai.abilitiesExplorer.fields JavaScript filter for adding read-only table columns to the Abilities Explorer page (#1181).
  • Clarified which parts of the plugin are stable Features vs Experiments, flag Experiments that run automatically in Settings > AI, and show when each runs, what it sends to your AI provider, and how many requests it makes from an info icon, the readme, and docs (#1190).
  • Featured Connector plugins (Mistral and Vercel AI Gateway) on Settings > Connectors, sorted by name, plus a wpai_featured_connectors filter to change the list (#1212).

變更

  • Update the wording in our capability check notice and change that from an error to a warning (#935).
  • Content Translation now translates any block that has rich text instead of just paragraphs and headings (#1009).
  • Move the Key Encryption experiment onto the new Secrets API feature plugin with backward-compatible fallback and migration (#1107).
  • The /<post>/feed/markdown/ no longer works and serves a 404 (#1121).
  • The core/content-query, core/settings-get, and core/users-query abilities now set the public meta flag added in WordPress 7.1 (#1122).
  • Summarize enabled and disabled experiments in the AI Status dashboard widget to keep it compact as more experiments are added (#1155).
  • Rebuilt the Abilities Explorer on DataViews and added three ai/v1 REST routes to retrieve and run abilities (#1181).
  • Replace the “Purge All Logs” button in the AI Request Logs screen with a control to delete logs older than 30 days, 90 days or 1 year, or all logs, and add a before_days argument to DELETE /ai/v1/logs (#1182).
  • The core/settings-get ability now matches WordPress core. Settings use their REST API names (for example, title instead of blogname), and a value that does not match its schema is left out. Breaking changes: Some schema fields have changed so ensure you are using the correct field names in your ability calls; the core/read-settings alias was also removed (#1206).
  • The core/users-query ability now matches WordPress core. The input is now where, fields, and pagination, and the output is always items and pagination. Users are looked up through the ids, slugs, or usernames lists in where, and lookup by email is no longer supported. Results are ordered by username, a filter that cannot be used is rejected, and a page after the last one returns an error. Breaking changes: Some schema fields have changed so ensure you are using the correct field names in your ability calls; the core/read-users alias was also removed (#1207).
  • The core/content-query ability now matches WordPress core. The input is now where, fields, and pagination, and the output is always items and pagination. Posts are looked up through the ids or slugs lists in where, which also requires the post type in type. Posts are filtered by author through where.author, with the author id or slug. Fields are nested, such as title.rendered. A filter that cannot be used is rejected, and a page after the last one returns an error. The author name and slug are only returned to users who may see the author. Only a show_in_abilities value of true exposes a post type. Breaking changes: Some schema fields have changed so ensure you are using the correct field names in your ability calls; the core/read-content alias was also removed (#1208).
  • The date and modified fields of core/content-query return site and gmt dates instead of local and gmt. Breaking change: Ensure you are using the correct field names in your ability calls (#1235).
  • The registered_date field of core/users-query is replaced by a registered object with site and gmt dates. Note that all 3 core abilities have now identical shape like these shipping in WordPress 7.2. Breaking change: Ensure you are using the correct field names in your ability calls (#1237).

移除

  • The deprecated core/read-settings ability alias. Use core/settings-get with REST API field names (for example, title instead of blogname). The alias could not keep the old field names or response shape after the sync with WordPress core (#1206).
  • The deprecated core/read-users ability alias. Use core/users-query with the new where, fields, and pagination input. The alias could not keep the old input and response shape after the sync with WordPress core (#1207).
  • The deprecated core/read-content ability alias. Use core/content-query with the new where, fields, and pagination input. The alias could not keep the old input and response shape after the sync with WordPress core (#1208).

修正

  • AI Request Logging now captures successful generations from providers that use a custom transport, not only those routed through the SDK HTTP transporter (#757).
  • Display a clear, context-aware error notice during Ability execution when the configured AI connector is pending approval (#759).
  • Remove the text generation capability check from the alt text generation WP-CLI command (#935).
  • Feature toggles are loaded with one database query instead of one per feature on sites without a persistent object cache (#1113).
  • Content Translation: Translate the blocks inside Post Content when “Show template” is enabled (#1116).
  • On multisite, network deactivation restores API keys on every site, and uninstall no longer deletes keys that were still encrypted (#1118).
  • The Markdown Feeds and Key Encryption flags are loaded together with the feature toggles instead of with one query each on every request (#1119).
  • The site-wide comment markdown feed lists comments instead of the post list (#1121).
  • Feed requests are no longer served the post’s Markdown by Accept negotiation or ?output_format=markdown (#1121).
  • Accept-header negotiation now follows the header’s quality values and order, so a request that prefers HTML gets HTML and one that prefers Markdown gets Markdown (#1125).
  • Ensure canceling meta description generation stops the loading state. Note the actual request won’t be cancelled but the result from that will be discarded (#1128).
  • Show an error when a Note fails to save, in place of a “suggestions added” message, and keep reviewing the other blocks when one fails (#1131).
  • Skip blocks that are removed, replaced or edited mid-translation (#1137).
  • Ensure all strings can be translated in the Abilities Explorer pages (#1146).
  • Ensure all strings can be translated in the Content Classification experiment (#1147).
  • A comment reply suggestion is no longer inserted into the reply form of a different comment (#1158).
  • Match existing terms whose names contain “&” or non-ASCII letters, and add them from suggestions (#1168).
  • A non-string term or parent in the AI response no longer fails the whole request when running content classification (#1171).
  • Add missing title attribute to singular alternate Markdown discovery link (#1177).
  • Show “Regenerate Reply” after a reply suggestion has been generated (#1179).
  • Checkbox icon color on the AI settings page (#1183).
  • Ensure success and error notices show on the AI Request Logs screen (#1184).
  • Key Encryption no longer causes a fatal error when stored API keys can’t be decrypted (#1186).
  • Suggest Reply no longer sends the title or excerpt of private or password-protected posts to the AI provider (#1192).
  • Comment Moderation no longer automatically analyzes pingbacks, trackbacks or block editor notes (#1192).
  • Uninstalling the plugin now removes Comment Moderation analysis data (#1192).
  • Omit Author meta line in Markdown feeds when a post has no assigned author (#1210).
  • Preserve line-level diff granularity for multi-line blocks like tables and quotes in Revision Change Summary (#1216).
  • Markdown Feeds now apply the core the_content_feed and the_excerpt_rss filters, matching RSS and Atom feed output (#1224).
  • Ensure the “Serve Markdown via the Accept header” setting can now be saved (#1229).
  • The Guidelines prompt no longer contains the HTML entity &#39; (#1230).
  • The permission error for a post now mentions image prompts instead of excerpts when running image prompt generation (#1234).

安全性

1.4.0 – 2026-10-05

新增

  • 執行留言審核時，會依每則留言的內容是否充實、是否符合該文章的主題，計算價值分數，提供介於 0 至 1 的相關性指標。這項分數會顯示在留言畫面的新欄位中（#681）。
  • 新增實驗功能 Markdown 資訊提供，在 /feed/markdown/ 提供 Markdown 資訊提供（適用於所有資訊提供情境），並透過 ?output_format=markdown 提供個別文章與頁面的 Markdown 版本（#855）。
  • 內容翻譯失敗時，使用者可手動重試（#941）。
  • 限制留言審核批次分析佇列的大小（#972）。
  • 新增嵌入向量的儲存與建立、讀取、更新、刪除（CRUD）層（新的 wpai_embeddings 資料表），可為每個向量一併記錄提供者與模型（#976）。
  • 為內容翻譯實驗功能新增 9 種預設目標語言（#986）。
  • 為內容翻譯能力新增指引分類支援（#987）。
  • 新增 Vector_Math 與 Vector_Ranker 類別，讓使用者能對向量執行相似度與排名查詢（#993）。
  • 為自動完成建議請求新增載入指示器（#1029）。
  • 將替代文字產生功能與圖片區塊的裝飾性圖片設定整合（#1031）。

變更

  • 將最低支援的 WordPress 版本提高至 7.0.3（#1001）。
  • 內容分類通知的錯誤訊息改為顯示具體的分類法標籤，不再使用籠統的用語（#916）。
  • 文章標題過短時顯示通知，並停用標題翻譯（#967）。
  • 將嵌入向量程式碼更新為
    PHP AI Client 的最新版本（#975）。
  • generate_embeddings 輔助函式現在必須指定
    嵌入模型。請傳入 model（模型 ID 或 ModelInterface
    執行個體）；若使用模型 ID，還須傳入提供該模型的 provider。這項
    變更會破壞相容性，已開始使用這個函式的人需要調整程式碼（#975）。
  • 將設定中的垂直刪節號圖示改為工具圖示（#1001）。
  • 將 core/read-content 與 core/read-users 能力重新命名為 core/content-query 與 core/users-query。舊名稱目前仍可作為已淘汰的別名使用（#1002）。
  • 將 gpt-image-2.5-flare 設為預設的 OpenAI 圖片產生模型（#1023）。
  • 將 core/read-settings 能力重新命名為 core/settings-get。舊名稱目前仍可作為已淘汰的別名使用（#1087）。

已淘汰

  • 淘汰 ai/get-post-details 能力。請改用 core/content-query 的單篇文章模式（#1002）。

移除

  • 移除已由 WordPress 核心處理、無須重複執行的網址驗證（#1011）。
  • 移除 AI 設定頁面頂端的 啟用 AI 開關（#985）。

修正

  • 確保內容調整長度互動視窗中的內容產生完成後，焦點會移至接受按鈕（#917）。
  • 核心重新驗證 API 金鑰時，將預設請求逾時時間從 5 秒延長至 30 秒。這是為了改善驗證請求逾時時，核心會刪除 API 金鑰的問題（#947）。
  • 將已啟用 SEO 外掛的偵測結果快取並設定存留時間（TTL），且在任何外掛啟用或停用時立即使快取失效，讓 Meta 描述的中繼資料鍵查詢不必在每次請求都重新掃描已啟用的外掛（#973）。
  • 改善內容翻譯的載入動畫，保留現有的文字與背景色彩（#977）。
  • 修正 @wordpress/dataviews 從私用 API 允許清單移除後，設定頁面在 Gutenberg 23.9 以上版本無法轉譯的問題（#989）。
  • 確保指引整合改從 wp_knowledge 文章類型讀取已發佈的指引資料，而不是已移除的 wp_guideline，讓指引可在最新版 Gutenberg 中使用。我們不會將舊文章類型與分類法的資料移轉至新版本，因此先前在舊版 Gutenberg 中設定過指引的人，需要重新新增這些指引（#988）。
  • 匯入核心加上前置詞的 PSR EventDispatcher，供內附的 EmbeddingBuilder 使用（#1005）。
  • 文字游標位於末尾時，為自動完成產生的文字預留空間（#1008）。
  • 更新設定頁面容器的 min-height，讓它涵蓋整個檢視區域的高度（#1013）。
  • 將 @wordpress/dataviews 套件更新為最新版本，修正可收合卡片的焦點樣式（#1035）。
  • AI 請求記錄現在會將提供者的 HTTP 錯誤回應（4xx/5xx）記錄為 error，並附上狀態碼與訊息，不再記錄為 success；Gemini 文字產生請求也不再被誤標為 Metadata（#1040）。
  • 行內留言表單處於快速編輯模式時，隱藏「建議回覆」按鈕（#1049）。
  • 修正區塊編輯器啟用「顯示範本」模式時，編輯備註只檢查範本區塊的問題（#1053）。
  • 啟用顯示範本時，將產生的摘要插入文章內容中（#1055）。
  • 將內容分類建議膠囊按鈕的焦點樣式與核心 Button 元件統一（#1060）。
  • 連接器核准功能不再因將核心的連接器金鑰檢查誤判為提供者外掛或 Gutenberg 發起的請求，而阻擋該檢查（#1070）。
  • 能力的輸入屬性列出多種類型時，能力總管不再當掉（#1073）。
  • 正規化內容時，保留 HTML 實體與短代碼之間的文字，同時仍對以實體編碼表示的標籤進行編碼（#1077）。
  • 避免啟用能力時，儲存一般設定會在 options.php 發生驗證錯誤（#1080）。
  • 確保 Markdown 資訊提供中的文章標題、網站說明及內容摘要不再含有 HTML 實體（#1086）。
  • 確保在資訊提供未註冊的情況下重建重新寫入規則後，Markdown 資訊提供不會出現 404 錯誤（#1110）。

安全性

  • 確保在允許批次審核留言之前，先檢查 moderate_comments 權限，再逐一檢查每則待審核留言的 edit_comment 權限（GHSA-pjw7-q94g-4q34）。
  • 強化 Playground 預覽發佈工作流程，改從觸發該流程的執行紀錄取得提取要求與提交的中繼資料（#978）。
  • 強化後台畫面的 JSON 資料解析（#991）。

較早的更新記錄可在 CHANGELOG.md 檔案中找到。

社群版外掛

這個外掛由社群人員開發及提供技術支援。 參與這個外掛專案

中繼資料

評分

4.2 星，滿分為 5 星

撰寫評分

查看全部使用者評論

技術支援

最近 2 個月解決的問題:

總計 6 個問題，已解決 4 個

檢視技術支援論壇

Image
Image