ImgX - See it. Prompt it.

AI prompt tool — hover any image, generate AI drawing prompts

ImgX - See it. Prompt it. What is ImgX? Hover any image - AI breaks down its visual grammar (subject / composition / lighting / color...) into ready-to-use AI drawing prompts. Two modes: sign in for free Cloud credits, or bring your own API key. Features • 🖱 Hover to analyze - works on any web image • ☁ Cloud free credits - sign in with any email (Gmail / Outlook / Yahoo / work email), 10 credits to start; no API key needed • 🔑 Bring your own key - per-provider configs, switch without losing keys; no middle server • 🧠 Smart routing - 8 templates auto-matched to image type • 🎨 Visual breakdown - 9 dimensions (subject / composition / lighting / color...) • 📊 Structured view - 12 fields, one-click copy • 📸 Capture & analyze - drag-select any area of the page • 📤 Share card - 1080px HD, glassmorphism • 📋 Batch queue - parallel + auto-retry Privacy - Two modes. BYOK (default): Key stored locally, sent only to your endpoint; images released after use; no server, no telemetry. Cloud (opt-in): email login, images proxied through the ImgX Worker to the model then immediately released - no image pixels stored; email + credit ledger kept in Supabase (Singapore) for billing and idempotency only. See privacy policy. Permissions • activeTab - Grants temporary access to the current tab only after you click ImgX's "Capture & analyze" button, to capture the visible area you select. The extension does not continuously access pages in the background, nor read or upload browsing history. • unlimitedStorage - Stores analysis task history and preprocessed image data from screenshot analysis locally, preventing screenshot data URLs from exceeding Chrome's default 10MB storage quota. All local; deletable from the extension's queue. • identity - Used only when you opt into Cloud mode and choose Google sign-in, to authenticate with Google and issue your ImgX session. Email-OTP sign-in does not use Google identity. No password is ever collected. • webNavigation - Used only during Google sign-in to capture the OAuth redirect (chromiumapp.org callback). The listener is removed automatically after a successful sign-in. • storage - Saves your settings (API key, endpoint, model), analysis history, queue, and optionally the Cloud login session (access/refresh token + email) when you opt into Cloud mode. All local data stays in chrome.storage.local. Needs - Either an OpenAI-compatible Vision Key (e.g. OpenAI / Gemini / Bailian / Zhipu - see settings panel) for BYOK mode, or just sign in with email for 20 free Cloud credits.You can receive 3 free credits for daily check-in. ———— ImgX - See it. Prompt it. ImgX 是什么？ 悬停任意图片，AI 拆解其视觉语法（主体 / 构图 / 光照 / 色彩……），生成中英双语 AI 绘图 Prompt。两种模式：登录领云端免费额度，或自带 API Key。 核心功能 • 🖱 悬停即分析 - 任意网页图片通用 • ☁ 云端免费额度 - 通用邮箱登录（QQ / 163 / Gmail / 企业邮箱），首账号赠 10 次；无需 API Key • 🔑 自带 Key - 各服务商独立保存，切换不丢；无中间服务器 • 🧠 智能路由 - 按图类型自动匹配 8 套模板 • 🌐 中英双语 - 中文 + 英文，9 大视觉维度 • 📊 结构化 - 12 字段，一键复制 • 📸 框选截图分析 - 拖选页面任意区域 • 📤 分享图 - 1080px 高清 • 📋 批量队列 - 并行 + 自动重试 隐私 - 两种模式。自带 Key（默认）：Key 仅存本地，只发往你的 endpoint；图片用完即释放；无服务器、无遥测。内置服务（可选）：邮箱登录，图片经 ImgX Worker 代理转发给模型后立即释放、不落盘；邮箱与额度流水存于 Supabase（新加坡），仅用于计费与幂等。详见隐私政策。 权限说明 • activeTab：用于用户主动点击 ImgX 的「截图分析」按钮后，临时访问当前活动标签页并截取用户框选的可见区域。扩展不会在后台持续访问网页，也不会读取或上传浏览历史。 • unlimitedStorage：用于仅在本机保存分析任务历史和截图分析产生的预处理图片数据，避免截图 data URL 超过 Chrome 默认的 10MB 存储配额。均为本地数据，可在扩展队列中删除。 • identity：仅在用户选择内置服务并使用 Google 登录时使用，用于向 Google 鉴权并签发 ImgX 会话。邮箱验证码登录不使用 Google identity。绝不收集密码。 • webNavigation：仅在 Google 登录时用于捕获 OAuth 回调（chromiumapp.org 重定向），登录成功后自动移除监听器。 • storage：保存用户设置（API Key、endpoint、模型）、分析历史、队列，以及选择内置服务时的云端登录会话（access/refresh token + 邮箱）。所有本地数据均存于 chrome.storage.local。 前提 - 自带 Key 模式需一个 OpenAI 兼容的 Vision Key（如 OpenAI / Gemini / 百炼 / 智谱，见设置面板）；或直接邮箱登录领 10 次云端免费额度，每日签到可领取 3 次免费额度。