Live Stream Interface Mockup

原始文档

GPT Image 2 · en

{
  "type": "Live Streaming Interface Model",
  "layout": {
    "main_feed": {
      "content": "A scene from a video game: a cowboy silhouette gazes at a tombstone bearing a Celtic cross at sunset.",
      "subtitle": "{argument name=\"game subtitle\" default=\"Arthur I gave you all I had.\"}",
      "ui_elements": [
        "The small circular map in the lower left corner"
      ]
    },
    "webcam": {
      "position": "Top left corner",
      "subject": "{argument name=\"webcam subject\" default=\"Sam Altman\"} His expression was focused and slightly sad.",
      "background": "A room with plants and a light-up lamp"
    },
    "stream_info": {
      "position": "bottom",
      "profile": {
        "name": "{argument name=\"streamer name\" default=\"SamAltman\"}",
        "verified": true,
        "avatar": "grayscale portrait of the anchor",
        "live_badge": "During the live stream"
      },
      "details": {
        "title": "{argument name=\"stream title\" default=\"first playthrough of RDR2 (no spoilers please!)\"}",
        "category": "{argument name=\"game title\" default=\"Red Dead Redemption 2\"}",
        "tags": [
          "English",
          "AMA",
          "OpenAI",
          "CEO",
          "AI"
        ]
      },
      "stats": {
        "viewers": "141,572",
        "uptime": "3:42:17"
      },
      "buttons": [
        "focus on",
        "subscription"
      ]
    },
    "chat_sidebar": {
      "position": "right side",
      "header": "Live chat",
      "messages_count": 17,
      "message_examples": [
        "spaghetti_code: Don't cry",
        "nicodemus88: We will all succeed.",
        "Quantumlance: A masterpiece",
        "souljaboytellem: Back pain (LUMBAGO)",
        "m4xw311: Riding the Highlands, Cowboy"
      ],
      "input_area": "Message input box with a purple \"Send\" button"
    }
  }
}

使用方法

1. 复制完整原始提示词，并保留其中的参数。
2. 按需求替换主体、文字和构图要求。
3. 在兼容的外部工具中使用，核对当前使用条件并检查结果。

开始前需了解

• 先确认是否需要参考图。标注兼容模型，不代表该示例由这个模型生成。
• 核对名称、数字与画面文字；不同模型版本的结果可能不同。