🎤 “把要读的字拍清楚，我就读给你听！”
这是一款纯粹、干净、充满温度的 Android 原生应用。本项目最初是为了帮助不识字的长辈而开发的，但它同样也是一个极简、高效、绝对安全的离线 OCR 工具。只需轻轻一拍，手机就能自动提取图片中的文字。
与市面上沉重、满是广告的文字识别工具不同，本项目追求的是极致的简单与安全：
- 智能逐行重排：内置坐标校准算法，强制将文字按真实的视觉阅读习惯（自上而下，从左往右）重新组合。解决了传统 OCR 乱序读取（如先读左半区、再读右半区）的痛点，让通知单、票据读起来逻辑通顺。
- 红点同步高亮 (Synchronized Visual Highlight)：播报时，图片上会实时跳动一个醒目的红色圆点，声音读到哪，红点指到哪，让长辈看得清、跟得上。
- 智能去噪：自动过滤 Emoji、乱码和线条。即使 AI 将图标误认为乱码，系统也会智能拦截，确保朗读内容清晰纯净。
- 即拍即得：没有复杂的配置和冗长的等待，识别速度极快。
- 一键复制：识别出的文字支持长按选中复制。无论是在复杂的背景中提取一段文字，还是整理文档，都能轻松胜任。
- 重听功能：配备醒目的“🎤 重听”按钮，没听清或没看清可以无限次反复重播。
- 100% 离线运行：无需网络，甚至未申请网络权限。所有文字识别和语音合成都在手机本地完成，不上传任何数据到云端，绝对保护隐私。
- 零广告、零付费：纯粹的公益开源项目，没有启动广告，没有会员弹窗。
- 自动清理存储：每次拍照和退出应用时，会自动清理所有临时图片缓存。它不仅保护你的隐私，还不占用你宝贵的存储空间。
- 开发语言：Kotlin
- AI 引擎：Google ML Kit (Text Recognition) - 高精度本地 OCR 引擎
- UI 实现：Android 原生 XML + 自定义 View (Canvas 绘制高亮)
- 语音系统：Android 原生 TextToSpeech (TTS)
- 最低兼容：Android 7.0 (API Level 24)
- 进入发布页面：点击 GitHub Releases 查看最新版本。
- 下载安装包：在最新的 Release 下找到
.apk后缀的文件并下载。 - 安装应用：安装时手机可能会提示“未知来源”，请选择“允许”或“继续安装”即可。
- 克隆本项目：
git clone https://github.com/robinji0/readinghelper.git - 使用 Android Studio 打开并 Build。
- 本项目已包含全套 GitHub Action 自动打包配置，提交代码后可手动触发 Release。
🎤 "Take a clear photo of the text, and I'll read it for you!"
This is a pure, clean, and warm Android native application. Originally developed to help elders who struggle with reading, it is also a minimalist, efficient, and absolutely secure offline OCR tool. With just one tap, it automatically extracts text from images.
Unlike complex and ad-filled OCR tools, this project pursues ultimate simplicity and security:
- Smart Line Reordering: Built-in coordinate calibration algorithm forces text into a natural reading order (top-to-bottom, left-to-right). This solves the "disordered reading" pain point of traditional OCR, making notifications and receipts logical to hear.
- Synchronized Visual Highlight: During playback, a prominent red dot moves across the image. Where the voice reads, the red dot points, helping elders follow along easily.
- Smart Denoising: Automatically filters emojis, gibberish, and lines. Even if the AI misidentifies an icon as a character, the system intelligently intercepts it to ensure clean audio.
- Instant Results: No complex configurations, extremely fast recognition speed.
- One-Click Copy: Extracted text supports long-press selection and copying. Perfect for extracting text from complex backgrounds or organizing documents.
- Replay Function: Equipped with a bright "🎤 Replay" button to listen again as many times as needed.
- 100% Offline: Works without internet; doesn't even request the
INTERNETpermission. All OCR and TTS are processed locally on the device, ensuring absolute privacy. - Zero Ads, Zero Fees: A pure non-profit open-source project with no splash ads or membership popups.
- Auto Storage Cleanup: Automatically clears all temporary image caches every time you take a photo or exit the app. It protects your privacy and saves storage space.
- Language: Kotlin
- AI Engine: Google ML Kit (Text Recognition) - High-precision local OCR engine.
- UI: Android Native XML + Custom View (Canvas for highlighting).
- Voice: Android Native TextToSpeech (TTS).
- Min SDK: Android 7.0 (API Level 24).
- Go to Releases: Click GitHub Releases for the latest version.
- Download APK: Find the
.apkfile (e.g.,ReadingHelper.apk) in the latest release. - Install: Send the APK to your phone and install it. (You may need to allow "Unknown Sources").
- Clone:
git clone https://github.com/robinji0/readinghelper.git - Build with Android Studio.
- Includes GitHub Actions for manual release builds.