ImgX

About

What is ImgX?

ImgX is a Chrome extension — hover any image, AI analyzes its visual grammar (subject, composition, lighting, color, materials...), generating bilingual AI drawing prompts in Chinese and English.

Key Features

🖱 Hover to analyze: Move your mouse over an image, a "Prompt" button appears. Click to start analysis
🧠 VL Model Powered: Built on multimodal VL models with 8 specialized prompt templates auto-routed by image type
🌐 Bilingual output: Natural Chinese paragraphs (150+ chars) + English paragraphs (80+ words), covering 9 visual dimensions
📊 Structured analysis: 12 structural fields (subject/action/details/environment/lighting/composition/style/color/material…), 3-tab view (zh / EN / JSON)
📋 One-click copy: Full-width copy button at the bottom for current tab content
📤 Export share card: Pure Canvas 2D 1080px HD share image with glassmorphism + gradient border
📚 History recall: Hover over previously analyzed images to restore results (zero API cost)
⚡ Batch queue: Enqueue multiple images, background engine processes in parallel, real-time progress panel
🔑 Your own key: Bring your own API Key, data goes directly to your endpoint, no intermediary server

8 Specialized Prompt Templates

ImgX's auto-classifier routes each image to the most suitable template: portrait (lighting + focal length), landscape (time + three-layer space), architecture (materials + perspective), product (lighting + shadows), UI design (platform + components), illustration (art style taxonomy), typography/poster (type + grid), and a general fallback.

Privacy by Design

API Key stored locally in your browser, sent only as a header to your configured endpoint
Image data used only for analysis, released immediately after processing
No intermediary server, no telemetry, no tracking
Full privacy policy at PRIVACY.md

Prerequisites

You need an OpenAI-compatible Vision API Key (recommended: SiliconFlow — free tier is sufficient for daily use).



ImgX 是什么？

ImgX 是一个 Chrome 扩展——鼠标悬停任意图片，AI 自动分析其视觉语法（主体、构图、光照、色彩、材质……），生成中英双语 AI 绘图 Prompt。

核心功能

🖱 悬停即分析：鼠标移到图片上，自动浮现「提示词」按钮，点击即开始分析
🧠 VL 模型驱动：基于 VL 多模态大模型，支持 8 套专属 Prompt 模板按图类型自动路由
🌐 中英双语：中文自然段落（≥150 字）+ 英文段落（≥80 词），覆盖 9 大视觉维度
📊 结构化分析：12 个结构字段（主体/动作/细节/环境/光照/构图/风格/色彩/材质…），中 / EN / JSON 三 Tab 切换
📋 一键复制：底部长条按钮，复制当前 Tab 内容
📤 导出分享图：纯 Canvas 2D 手绘 1080px 高清分享卡，玻璃质感 + 渐变描边
📚 历史可召回：已分析过的图悬停自动识别，唤回历史结果（不消耗 API）
⚡ 批量队列：多图分析入队，后台引擎并行跑，队列面板实时看进度
🔑 自有 Key：填入自己的 API Key，数据直连 API，无中间服务器

8 套专属 Prompt 模板

ImgX 内置自动分类器，按图类型路由到最合适的 Prompt 模板：人物（肖像/全身）、风景（时段+空间三层）、建筑/室内（材质+透视）、商业产品（灯光+阴影）、UI 界面（平台+组件）、插画/二次元（画风+线条）、海报/排版（字体+网格），以及通用兜底模板。

隐私设计

API Key 仅存储在浏览器本地，仅作为请求头发送到你配置的 endpoint
图片数据仅用于分析，处理完成即释放
无中间服务器、无遥测、无追踪
完整隐私政策见 PRIVACY.md

使用前提

你需要一个 OpenAI 兼容的 Vision API Key（推荐 硅基流动，免费额度足够日常使用）。