Prompt library

The prompt library (titled "提示词中心" in the app) collects proven descriptions that you can reuse in Image, Video, and Canvas. Administrators maintain the records; creators browse, copy, and save them as assets.

Browsing and filtering

• Search matches the prompt title.
• Category is a dropdown built from currently enabled prompts (administrators type them, for example 摄影、海报、角色).
• Tags support multiple selections.

The list is a card grid that loads as you scroll (20 per page) and ends with "已显示全部结果" (or "正在加载更多…" while loading). Order comes from the administrator's sort value; the interface has no sorting control.

Each card shows a cover or a text placeholder, the title, the update time, a ~120-character summary, and up to four tags, with "复制" and "存为素材" actions.

Viewing and using

• View details: clicking a card opens a dialog with the full text, category and tags, and the created and updated timestamps; its buttons are "复制提示词" and "加入我的资产".
• Copy puts the text on the clipboard and reports "提示词已复制".
• Save as asset writes it into My assets as a text asset whose source is the prompt category, reporting "已加入我的资产".
• Bring it into generation: the "提示词库" button in the Image or Video composer opens the "选择提示词" dialog; clicking "使用" fills the composer, and you adjust parameters for the current task. "从提示词库选择" on a canvas node works the same way.

Administrator maintenance

Administrators maintain records in the "提示词库" tab of the admin console. Fields: title, content, category, tags, preview content, source URL, sort order, and status.

• Enable / disable: disabling hides the prompt from users ("停用后用户侧提示词库将不再展示该提示词。"); anything already copied is unaffected.
• Sort order: lower values come first; the default is 1000.
• Preview content: Markdown, used to show reference images and notes in the detail dialog.
• The source URL is not shown to users yet.

Other console capabilities are described in Admin console.