ReadAny

codedogQBY活跃维护

跨平台电子书阅读器，内置本地向量库，支持语义搜索、基于 RAG 的书籍问答、笔记、TTS 朗读和 WebDAV 同步，可接入 OpenAI、Anthropic、Ollama 等模型。

AI-powered cross-platform e-book reader with semantic search, RAG chat, local vector store, notes, TTS, and WebDAV sync.

v1.3.6发布于 2026-08-16 07:35 UTC2,962star +13 / 7 天244fork同步于 2026-11-10 05:41 UTC

电子书阅读器语义搜索 RAG问答本地向量库语音合成

对比 GitHub 官网 ai anthropic ebook-reader embeddings epub-reader

怎么上手

由 README 提炼，执行前请确认

原文

brew tap codedogQBY/readany
brew install --cask readany

硬件要求按 README 说明

README 未说明显存或 GPU 要求；桌面端基于 Tauri，提供 macOS（Apple Silicon/Intel）、Windows、Linux 安装包，移动端支持 iOS 与 Android。

支持 Apple Silicon

下载：源码 zip·GitHub Releases（安装包）本站不镜像安装包

项目说明

由 README 自动翻译提炼

没人工点评？写下你的经验

本地优先的 AI 电子书阅读器，用 RAG 对话和语义搜索把读过的书变成可检索的私人知识库

跨平台电子书阅读器，内置本地向量库，支持语义搜索、基于 RAG 的书籍问答、笔记、TTS 朗读和 WebDAV 同步，可接入 OpenAI、Anthropic、Ollama 等模型。

核心能力

用向量检索加 BM25 混合搜索，按语义找书中概念
基于当前书籍和选中段落做 RAG 对话并定位来源
本地嵌入与本地向量库，书籍笔记可离线使用
支持 OpenAI、Claude、Gemini、Ollama、DeepSeek 等模型
内置摘要、概念解释等 Skills，可自建自定义技能

容易踩的坑

AI 功能需在设置里自行填入各家 API Key，不配置则无法对话；移动端开发必须用 expo-dev-client 构建，Expo Go 无法加载原生模块；开发环境需 Node.js ≥18、pnpm ≥9 和 Rust，移动端另需 Xcode 或 Android Studio。

内容由模型根据 README 生成，可能不完整或不准确，以官方文档为准。看原文 README

最近发版

5 个版本v1.3.6

全部 release

v1.3.6ReadAny v1.3.6最新正式版2026-08-16 · 2026-08-16 07:35 UTC
ReadAny v1.3.6 来啦。这一版重点完善了 桌面端与移动端的 RAG / 知识检索能力，新增 PDF 深色与 Sepia 主题适配，并集中改善 AI 对话、引用跳转、移动端输入以及 EPUB / PDF 阅读体验。

说明：本版本说明只列用户可感知的新增、优化和修复。开发过程中的内部验证、联调和边界问题收敛不单独列出。

新功能
- PDF 支持深色与 Sepia 主题：桌面端和移动端的 PDF 页面可跟随全局主题，并逐页识别内容，尽量避免对照片或深色页面错误反色
- 新增 Atlas Cloud AI 服务商：可在 AI 模型设置中配置和使用 Atlas Cloud
- AI 对话显示选中文本原句：从阅读器选中文字并进入 AI 对话后，可在顶部直接对照原文提问
- 移动端补全本地与远程向量检索：支持远程 Embedding，并为中文和英文提供本地模型选择
优化
- RAG 检索更完整可靠：统一桌面端和移动端的查询向量配置，校验书籍向量来源与维度，减少模型或服务切换后检索结果异常
- AI 工具调用与引用跳转更稳定：工具参数兼容性、超时处理和状态展示更可靠；缺少精确位置的引用也会尝试根据章节和原文恢复跳转
- AI 对话反馈更清楚：超长输入会提示分段提问，输出达到长度限制或工具调用未完成时不再把残缺回答显示为正常完成
- 移动端输入体验更完善：设置、首次引导、笔记、书籍详情、新建分组等输入区域会自动避让键盘
- 阅读器书籍样式兼容性更好：书籍字体设置可正确作用于正文，同时尊重书籍对行内图片高度的显式设置，并继续限制大图溢出
- 阅读器交互更稳：优化 iOS 长按选词、划线和翻页的冲突，链接点击、目录跳转与正常点按翻页之间的判断更准确
修复
- 修复 iOS 选词或划线结束后可能误触翻页，以及选中状态下翻页行为不稳定的问题
- 修复点击 EPUB 目录或正文链接后可能闪回原页、跳到上一章的问题
- 修复部分 EPUB 缺少样式表链接时书籍样式完全丢失的问题
- 修复部分 EPUB 行内小图标被异常放大，以及行内标记式注脚无法预览或跳转的问题
- 修复 macOS Retina 屏幕上 PDF 文本层错位、跨行选中和复制不准确的问题
- 修复移动端 PDF 在深色主题下出现色差、白边等显示问题
- 修复 AI 聊天中超长连续文本无法换行、深色模式下 Markdown 可读性较差的问题
- 修复部分 AI 工具参数无法生成 Schema、引用未登记或点击后无法跳转的问题
- 修复 Android TTS 音色选择器崩溃，以及小米 MiMo TTS 偶尔无法继续朗读的问题
- 修复卸载外部 AI / CLI 接入后残留配置或反馈不明确的问题
English Release Notes

ReadAny v1.3.6 improves RAG and knowledge retrieval across desktop and mobile, adds dark and Sepia themes for PDFs, and delivers a focused set of improvements to AI chat, citation navigation, mobile input, and EPUB / PDF reading.

Note: these release notes only list user-facing additions, improvements, and fixes. Internal validation, integration, and edge-case cleanup are not listed separately.

New Features
- Added dark and Sepia themes for PDFs on desktop and mobile, with per-page content detection to avoid incorrectly inverting photo-heavy or already-dark pages
- Added the Atlas Cloud AI provider to AI model settings
- Added selected-text context to AI chat so the original passage remains visible when asking about a selection from the reader
- Completed local and remote vector retrieval on mobile, including remote embeddings and local model options for Chinese and English
Improvements
- RAG retrieval is more complete and reliable with aligned query-embedding configuration across desktop and mobile, embedding provenance validation, and safer handling of model dimension changes
- AI tool calls and citation navigation are more stable with more compatible parameters, clearer timeouts and status, and fallback location recovery from chapter text
- AI chat feedback is clearer: oversized input asks users to split the request, while token-limited or incomplete tool responses are no longer shown as successfully completed answers
- Mobile input handling is more reliable across settings, onboarding, notes, book details, and new-group flows by keeping fields visible above the keyboard
- Book styling compatibility is improved: book-font settings now apply correctly to body text, explicit inline-image heights are respected, and large images remain constrained to the viewport
- Reader interactions are more stable around iOS text selection, highlighting, page turns, links, and table-of-contents navigation
Fixes
- Fixed accidental page turns after ending text selection or highlighting on iOS, plus selected-state page-turn inconsistencies
- Fixed EPUB TOC and content links occasionally flashing back or navigating to the previous chapter
- Fixed missing book styles in EPUB files whose chapters do not explicitly link their manifest stylesheets
- Fixed oversized inline EPUB icons and inline-marker footnotes that could not be previewed or opened
- Fixed misaligned PDF text layers, multiline selection, and inaccurate copy behavior on macOS Retina displays
- Fixed selected mobile PDF dark-theme color mismatch and thin white edges
- Fixed long unbroken AI chat messages overflowing and Markdown readability iss
v1.3.5ReadAny v1.3.52026-07-08 · 2026-07-08 07:47 UTC
查看更新说明
ReadAny v1.3.5 来啦。这一版重点带来 ReadAny CLI / 外部 AI 接入能力，同时升级了 TTS 语音服务配置，并继续改善 AI 问答、阅读器、移动端朗读和同步体验。

说明：本版本说明只列用户可感知的新增、优化和修复。开发过程中为新功能做的内部校验、联调、验收和边界问题收敛，不单独作为“优化”或“修复”列出。

新功能
- 新增 ReadAny CLI：提供命令行入口，可用于读取书库信息、检索知识、导出笔记 / 知识、检查环境、管理配置等高级工作流
- 新增 MCP / 外部 AI Agent 接入能力：支持把 ReadAny 的书库、阅读上下文、笔记、高亮、知识搜索等能力接入外部 AI 工具和 Agent 客户端
- 新增外部 AI 访问设置（桌面端）：新增外部 AI / MCP 配置页，支持生成配置、查看可用状态、诊断连接和查看审计日志
- 新增 EPUB Draft 工作区（桌面端，高级功能）：支持创建 EPUB 草稿、编辑章节与元信息、查看差异并导出修改后的 EPUB
- 新增更灵活的 TTS 服务配置：TTS 设置支持按服务商维护配置，为后续接入更多语音引擎打基础
- 新增 Xiaomi TTS / 云 TTS 相关支持：补充小米 TTS 配置、声音列表、Base URL 与试听流程
- 新增系统 TTS 原生合成能力（移动端）：移动端补充系统 TTS 原生模块，提升本机朗读兼容性
优化
- 外部 AI 接入更易配置：桌面端提供更清晰的外部 AI / MCP 设置入口，方便生成配置、检查状态并定位连接问题
- TTS 设置体验更清晰：朗读服务、声音、试听反馈和错误提示更集中，后续接入更多语音引擎时也更容易管理
- TTS 播放更稳定：暂停、继续、取消、预取，以及切换音色 / 语速 / 音调后的播放体验更顺滑
- AI 问答连续性更好：AI 正在回答时切回阅读页、切到其他页面后再回来，正在生成的问答会继续保留
- 阅读上下文衔接更稳：桌面端 AI 引用回到阅读器、阅读上下文同步等链路更可靠
- 移动端阅读体验更完善：竖排 EPUB、固定版式 / 图片页、目录跳转、滚动跨章节、字体大小和选区操作等细节更稳定
- PDF / EPUB 阅读体验更稳：桌面端 PDF 控件、缩放、分页跳转，以及 EPUB 图片资源处理更可靠
- 向量化与知识检索准备更充分：章节结构识别、远程 embedding 请求、PDF/EPUB 章节兜底和大文件处理更稳
- 同步和统计刷新更可靠：WebDAV 目录处理、同步路径、本地文件路径、书籍元信息变化后的统计刷新更稳定
修复
- 修复 AI 回答过程中离开聊天页 / 阅读页后再回来，当前问答可能丢失的问题
- 修复云 TTS 暂停、取消、提前跳句、进度不同步等播放稳定性问题
- 修复播放中切换朗读引擎、音色、语速或音调后，部分场景下可能出现旧音频残留或重复播放的问题
- 修复移动端从后台回到前台后，部分场景下 TTS 状态不同步的问题
- 修复移动端部分 EPUB 阅读场景下竖排、固定版式、图片页、RTL 翻页、CFI 跳转等兼容问题
- 修复桌面端部分 PDF 控件、页面滚动、缩放和阅读位置保存相关问题
- 修复部分 EPUB 图片资源、脚本处理、首屏分页设置应用相关问题
- 修复 WebDAV 目录大小写、重复 MKCOL、同步后本地文件路径等同步兼容问题
- 修复 Microsoft 中文翻译目标语言映射问题
English Release Notes

ReadAny v1.3.5 introduces ReadAny CLI / external AI access, upgrades TTS provider configuration, and improves AI chat, reader, mobile read-aloud, and sync experiences.

Note: these release notes only list user-facing additions, improvements, and fixes. Internal validation, integration, acceptance, and edge-case cleanup made while developing new features are not counted as standalone improvements or fixes.

New Features
- Added ReadAny CLI: a command-line entry point for library inspection, knowledge search, notes / knowledge export, environment checks, configuration management, and advanced workflows
- Added MCP / external AI Agent access: expose ReadAny library data, reading context, notes, highlights, and knowledge search to external AI tools and Agent clients
- Added desktop external AI access settings: generate MCP configs, check readiness, diagnose setup, and review audit logs from the app
- Added desktop EPUB Draft workspace as an advanced feature: create EPUB drafts, edit chapters and metadata, inspect diffs, and export updated EPUB files
- Added more flexible TTS provider configuration: TTS settings now support provider-based profiles, preparing the app for more voice engines
- Added Xiaomi TTS / cloud TTS support: added Xiaomi TTS configuration, voice options, Base URL handling, and preview flow improvements
- Added native system TTS synthesis on mobile for better device-level read-aloud compatibility
Improvements
- External AI access is easier to configure with a clearer desktop settings entry for generating configs, checking readiness, and diagnosing connection issues
- TTS settings are clearer with more focused provider selection, voice selection, preview feedback, and error states
- TTS playback is more stable around pause, resume, cancellation, prefetching, and playback after voice / rate / pitch changes
- AI chat continuity is improved: in-progress AI answers now survive leaving and reopening chat / reader screens
- Reading context handoff is more reliable when navigating from desktop AI citations back into the reader
- Mobile reading is more polished for vertical EPUB, fixed-layout / image-only pages, TOC navigation, cross-chapter scrolling, font sizing, and selection actions
- PDF / EPUB reading is more stable around desk
v1.3.4ReadAny v1.3.42026-06-09 · 2026-06-09 16:40 UTC
查看更新说明
ReadAny v1.3.4 来啦，这一版是一个偏修复与体验打磨的小版本，主要围绕 移动端图表查看、AI 输入框键盘适配、阅读器翻页 / 目录跳转、高亮操作 做了更新。

🚀 新功能
- 新增移动端图表全屏查看：AI 生成的思维导图和 Mermaid 图表现在可以进入全屏查看，更适合阅读复杂结构图
- 增强移动端思维导图手势浏览：支持在图表内更自然地拖拽、缩放，查看大图时不再局限于当前小窗口
✨ 优化
- 优化移动端思维导图触摸体验：改善 WebView 内部拖拽、双指缩放与页面滚动之间的手势冲突
- 优化移动端 Mermaid 图表展示：补充全屏查看能力，并改善图表在移动端的查看体验
- 优化移动端 AI 输入框键盘避让：调整底部输入区域、Tab 安全区域与键盘高度的处理，减少输入框被键盘遮挡的问题
- 优化思维导图 SVG 导出：导出的 SVG 会带上 XML 声明、SVG / XHTML 命名空间、内联样式与更安全的文件名，方便分享和在外部应用中打开
- 优化桌面端已有高亮的操作体验：点击已有高亮时，也会展示和选中文本一致的操作浮层，带笔记的高亮也能正常处理
- 更新微信群二维码
🐛 修复
- 阅读器导航：
  
  修复桌面端阅读器在新版阅读引擎后，左右点击翻页、键盘翻页、目录跳转可能全部失效的问题
  
  修复 Foliate overlayer 缺少 hole API 导致的 clearHole is not a function 崩溃
- 移动端图表与输入：
  
  修复移动端首次全屏查看思维导图 / Mermaid 图表时，安全区域计算不准确的问题
  
  修复移动端 AI 输入框在聚焦、发送消息或键盘弹出后位置 / 高度异常的问题
  
  修复移动端思维导图在 WebView 内拖拽、缩放不够稳定的问题
- 高亮操作：
  
  修复桌面端点击已有高亮时操作浮层不出现或定位不准确的问题
  
  修复带笔记高亮在点击后不容易继续编辑、删除或操作的问题
English Release Notes

ReadAny v1.3.4 is a focused patch release for mobile diagram viewing, AI input keyboard handling, reader navigation, and highlight actions.

🚀 New Features
- Added fullscreen diagram viewing on mobile: AI-generated mindmaps and Mermaid diagrams can now be opened fullscreen, making complex diagrams easier to inspect
- Improved mobile mindmap gesture browsing: mindmaps now support more natural dragging and zooming inside the diagram view
✨ Improvements
- Improved mobile mindmap touch handling: reduced conflicts between WebView gestures, diagram panning / zooming, and page scrolling
- Improved mobile Mermaid viewing with fullscreen support and a better mobile inspection experience
- Improved mobile AI input keyboard avoidance by refining bottom input layout, tab safe-area handling, and keyboard height behavior
- Improved mindmap SVG export with XML declaration, SVG / XHTML namespaces, inline styles, and safer filenames for sharing and opening in external apps
- Improved desktop existing-highlight actions: clicking an existing highlight now shows the same action popover as selected text, including highlights with notes
- Updated the WeChat group QR code
🐛 Fixes
- Reader navigation:
  
  Fixed a desktop reader regression where left / right tap navigation, keyboard page turns, and TOC jumps could all stop working after the reading engine update
  
  Fixed the clearHole is not a function crash caused by missing Foliate overlayer hole APIs
- Mobile diagrams and input:
  
  Fixed incorrect safe-area calculation when opening mindmaps / Mermaid diagrams fullscreen for the first time
  
  Fixed mobile AI input position / height issues after focusing, sending messages, or opening the keyboard
  
  Fixed unstable mindmap dragging and zooming inside the mobile WebView
- Highlight actions:
  
  Fixed desktop existing-highlight taps not showing the action popover or placing it incorrectly
  
  Fixed highlights with notes being harder to edit, delete, or act on after tapping
完整变更对比 / Full changelog: https://github.com/codedogQBY/ReadAny/compare/v1.3.3...v1.3.4

下载 / Downloads: 见下方 Assets。桌面端会通过更新通知自动提示升级。See Assets below. Desktop builds will also show the update through the in-app updater. iOS builds are distributed through EAS / App Store Connect / TestFlight.
v1.3.3ReadAny v1.3.32026-06-07 · 2026-06-07 15:33 UTC
查看更新说明
ReadAny v1.3.3 来啦，这一版主要围绕 书籍详情、连续滚动阅读、AI 引用与兜底、同步稳定性、TTS 听书体验 做了一轮更新。

🚀 新功能
- 新增书籍详情 / 元信息编辑：支持在桌面端与移动端维护书名、作者、标签、评分、简介、书评等信息
- 新增连续滚动阅读模式：阅读器支持更接近自然上下滚动的连续阅读体验，并持续优化章节衔接
- 新增 AI 未索引书籍兜底探索：没有向量化的书籍也能通过原文探索尽量回答问题，并生成可追溯引用
- 新增导入后自动向量化相关流程：为后续“导入即索引”的 AI 阅读体验打基础
- 新增高亮擦除操作：点击高亮后可以直接删除高亮
- 新增默认高亮颜色记忆：选择过的高亮颜色会被记住，减少重复选择
- 新增自定义 S3 同步目录：S3 同步也可以像 WebDAV 一样指定目录
- 新增移动端缓存清理入口
✨ 优化
- 书籍详情交互优化：桌面端与移动端分别做了更适合当前平台的展示与编辑体验
- 滚动阅读体验优化：改善滚动模式的章节切换、页边距、滚动条样式，以及桌面端点击翻页规则
- AI 引用流程更稳：AI 回答前会更主动生成引用；引用工具调用也会正常展示，便于理解 AI 正在做什么
- 未索引书籍引用定位优化：改进兜底引用到书内位置的映射，降低引用跳转不准的问题
- 同步进度更清晰：文件同步的上传、下载与进度反馈更准确
- WebDAV / S3 同步体验优化：减少移动端重复创建目录请求，并改善桌面端 WebDAV 兼容性
- TTS 播放控件体验优化：播放、暂停等按钮的状态文案和可访问性标签更清楚
- 移动端自定义字体导入优化：本地字体文件导入与启用更稳定
- README 支持格式说明更新
🐛 修复
- 阅读与标注：
  
  修复注音文本可能被一起加入高亮或笔记的问题
  
  修复选中文本悬浮工具栏可能超出阅读器边界的问题
  
  修复 PDF 阅读时中间区域点击控制不准确的问题
  
  修复脚注点击直接跳转的问题，改为在当前位置展示提示气泡，并处理边缘遮挡
  
  修复部分场景下阅读进度越界或新书进度展示不准确的问题
  
  修复桌面端滚动模式相关改动后，非滚动模式点击翻页位置可能错乱的问题
- AI / 向量化 / 翻译：
  
  修复未索引书籍 AI 引用顺序、位置映射和跳转不稳定的问题
  
  修复部分 AI 工具调用失败后缺少失败态、界面一直 loading 的问题
  
  修复移动端通过导出码导入 AI 配置后，部分凭据没有正确保存的问题
  
  优化中文现代释义提示词，减少古文被直接翻回原文或被过度解释的问题
- 同步与导入：
  
  修复同步后已删除分组可能被重新恢复的问题
  
  修复移动端下载书籍后文件尚未准备好就打开的问题
  
  修复 WebDAV 桌面端部分配置无法连接的问题
  
  修复移动端 WebDAV 在部分服务上重复 MKCOL 导致失败的问题
- TTS / 听书：
  
  修复 DashScope 朗读进度与实际音频不同步的问题
  
  修复 DashScope 暂停 / 继续后可能跳段、重复请求或音频混乱的问题
  
  修复朗读时可能出现重复句子的问题
  
  修复滚动模式下 TTS 起读位置可能不是当前视野第一句的问题
  
  修复桌面端阅读器文本识别在阅读引擎更新后失效的问题
  
  修复 TTS 读出脚注标记的问题
- 移动端与桌面端稳定性：
  
  修复阅读统计小时显示与插值问题
  
  修复移动端清理缓存、字体导入、书籍下载等细节问题
  
  清理被错误跟踪的 __pycache__ 文件
English Release Notes

ReadAny v1.3.3 focuses on book details, continuous scrolling, AI citations and fallback reading, sync reliability, and TTS playback improvements.

🚀 New Features
- Added book details / metadata editing: edit title, author, tags, rating, summary, reviews, and more on desktop and mobile
- Added continuous scroll reading mode: a more natural vertical reading experience with smoother chapter continuity
- Added AI fallback exploration for non-indexed books: AI can explore original book content and produce traceable citations even before vectorization
- Added import-time auto-vectorization workflow groundwork for a smoother “import and index” AI reading experience
- Added highlight erase action: delete highlights directly after tapping them
- Added remembered default highlight color to reduce repeated color selection
- Added custom S3 sync folder support, bringing S3 closer to the WebDAV directory workflow
- Added mobile cache clearing entry
✨ Improvements
- Improved book details UX with platform-specific desktop and mobile layouts
- Improved continuous scroll reading: better chapter transitions, margins, scrollbar styling, and desktop tap navigation behavior
- More reliable AI citation flow: AI now generates citations more proactively before final answers, and citation tool calls are visible in the chat
- Improved fallback citation mapping for non-indexed books, reducing inaccurate citation jumps
- Clearer sync progress with more accurate upload, download, and transfer feedback
- Improved WebDAV / S3 sync experience: reduced redundant mobile directory creation and improved desktop WebDAV compatibility
- Improved TTS playback controls with clearer state labels and accessibility labels
- Improved mobile custom font import reliability
- Updated README supported format documentation
🐛 Fixes
- Reading and annotations:
  
  Fixed ruby / pinyin annotation text leaking into highlights or notes
  
  Fixed the selection popover sometimes appearing outside the reader bounds
  
  Fixed inaccurate center-tap controls when reading PDFs
  
  Replaced direct footnote jumps with in-place tip bubbles and fixed edge clipping
  
  Fixed reading prog
v1.3.2ReadAny v1.3.22026-05-30 · 2026-05-30 17:01 UTC
查看更新说明
ReadAny v1.3.2 来啦，这一版主要围绕 移动端上架准备、阅读体验、注音、多语言、同步导入稳定性 做了一轮比较扎实的更新。

🚀 新功能
- 新增 UMD 格式电子书支持：可以导入并阅读 .umd 电子书
- 新增拼音注音功能：阅读中文内容时可开启拼音注音，支持桌面端与移动端
- 新增多语言界面：新增西班牙语、日语、韩语、法语，并补充繁体中文；同时清理了一批硬编码中英文文案
- 新增维吾尔语翻译目标语言
- 阅读器支持 PageUp / PageDown 翻页
- 反馈入口新增未读红点：有回复时更容易看到
- 移动端加入备案信息展示：在“我的”页底部展示 APP 备案号，方便应用商店审核与合规
✨ 优化
- 移动端章节切换更快：减少跨章节切换等待，去掉不必要的淡出延迟
- 阅读翻页更干脆：禁用部分页面切换动画，降低阅读时的等待感
- 移动端手势更稳定：横向翻页与下拉书签手势增加方向锁定，减少误触
- 字体导入不再限制 20MB
- WebDAV 导入体验优化：记住“连接另一个 WebDAV”的配置，带路径的 WebDAV 地址也能正确列目录
- 书架卡片对齐优化：无作者信息时卡片高度保持一致
- 自动向量化进度更清晰：导入后自动向量化会显示进度；重复索引前改为应用内确认弹窗
🐛 修复
- 阅读与标注：
  
  修复标注、高亮、笔记、书签列表没有按书中位置排序的问题
  
  修复移动端切换章节后，再切回上一章时高亮/笔记没有重新加载的问题
  
  修复 AI 获取上下文时拿到章节开头而不是当前可见页面的问题
  
  修复首次打开部分书籍时可能卡在空章节的问题
  
  修复隐藏标签页更新阅读样式可能导致界面卡住的问题
- 注音与 TTS：
  
  修复关闭注音后拼音文本可能残留到正文的问题
  
  修复 TTS 会把拼音注音读出来的问题
  
  修复没有块级元素时 TTS 找不到朗读内容的问题
  
  优化注音字典加载与覆盖率
- 同步与导入：
  
  修复 Windows WebDAV 同步遇到非 ASCII 响应头的问题
  
  修复 WebDAV 限流/401 场景下同步失败率较高的问题，并降低并发
  
  修复自建 S3 / rclone 场景下 path-style 兼容问题
  
  修复 WebDAV 导入路径重复问题
- AI / 向量化 / 翻译：
  
  修复 Windows 自定义 AI 端点测试连接和获取模型时的 CORS 问题
  
  自定义 AI Base URL 未填写 http:// 或 https:// 时自动补全
  
  向量化批量请求更小，并对 4xx 错误按 chunk 重试，减少整批失败
  
  修复 Microsoft 翻译英文到中文时可能返回原文的问题
- 移动端与桌面端稳定性：
  
  修复移动端聊天历史侧边栏快速开关时卡在半途的问题
  
  修复 Android TrackPlayer 重复初始化报错
  
  修复 iOS 开发环境冷启动偶发 TurboModule 红屏
  
  修复桌面端拖拽导入书籍失效的问题
  
  修复反馈日志过大导致提交失败的问题
English Release Notes

ReadAny v1.3.2 focuses on mobile App Store readiness, reading experience improvements, pinyin annotations, localization, and more reliable sync/import flows.

🚀 New Features
- Added UMD ebook support: import and read .umd books
- Added pinyin annotations: show pinyin while reading Chinese content, available on desktop and mobile
- Added more interface languages: Spanish, Japanese, Korean, French, plus Traditional Chinese; also cleaned up hard-coded Chinese/English strings
- Added Uyghur as a translation target language
- Added PageUp / PageDown page navigation in the reader
- Added unread indicator for feedback replies
- Added mobile app filing information: the app filing number is now shown at the bottom of the Me page for compliance and store review
✨ Improvements
- Faster chapter switching on mobile: reduced waiting time and removed unnecessary fade-out delay
- Snappier page turns: disabled some page transition animations to make reading feel more immediate
- More stable mobile gestures: added direction locking for horizontal page turns and pull-to-bookmark gestures to reduce accidental triggers
- Removed the 20 MB limit for font imports
- Improved WebDAV import experience: remembers the “connect another WebDAV” configuration and correctly lists directories when the WebDAV URL includes a path
- Improved bookshelf card alignment: card height stays consistent when author metadata is missing
- Clearer auto-vectorization progress: progress is shown after import, and duplicate indexing now uses an in-app confirmation dialog
🐛 Fixes
- Reading and annotations:
  
  Fixed highlights, notes, annotations, and bookmarks not being sorted by their position in the book
  
  Fixed highlights and notes not reloading on mobile after switching to another chapter and then returning
  
  Fixed AI context sometimes using the beginning of the chapter instead of the currently visible page
  
  Fixed some books getting stuck on an empty chapter when first opened
  
  Fixed a case where updating reading styles in a hidden tab could freeze the UI
- Pinyin annotations and TTS:
  
  Fixed pinyin text sometimes remaining in the body after annotations were turned off
  
  Fixed TTS reading pinyin annotations aloud
  
  Fixed TTS failing to find readable content when a chapter had no block-level elements
  
  Improved pinyin dictionary loading and coverage
- Sync and import:
  
  Fixed Windows WebDAV sync issues caused by non-ASCII response

用户点评

登录用户的真实使用记录，可按版本和硬件筛选

共 0 条

还没有人点评。第一个说出真实体验的人最有价值。

README 原文

未翻译

在 GitHub 打开

🚀 v2.0 Update: Mobile apps (iOS/Android) now available! See Mobile section below.

AI-Native Reading Workflow

Chat with your library - Ask questions about the current book or selected passages, with answers grounded in your reading context
Find ideas semantically - Hybrid vector retrieval + BM25 search helps you find concepts even when the exact keyword is missing
Keep knowledge private - Local embeddings and a local vector store keep your books, highlights, and notes offline-capable
Use your preferred model - Connect OpenAI, Claude, Gemini, Ollama, DeepSeek, or custom-compatible providers

Why ReadAny?

Problem	Traditional Readers	ReadAny
Search content	Keywords only	Semantic search that understands your intent
Ask questions	Find answers yourself	AI answers directly + locates sources
Take notes	Manual copy-paste	Select to highlight, one-click export
Knowledge management	Scattered notes	Unified management, multi-format export
Privacy	Upload to cloud	Local vector store, fully offline capable

Comparison with Alternatives

Feature	ReadAny	Calibre	KOReader	Apple Books
AI Chat	✅	❌	❌	❌
Semantic Search (RAG)	✅	❌	❌	❌
Local Vector Store	✅	-	-	❌
TTS (Text-to-Speech)	✅	❌	Limited	Limited
Reading Stats	✅	❌	❌	Limited
WebDAV Sync	✅	❌	❌	❌
Skills System	✅	❌	❌	❌
Format Support	10+	15+	10+	2
Note Export	5 formats	Limited	Limited	Limited
Open Source	✅	✅	✅	❌

Screenshots

Hero

AI-Powered Chat

Notes & Highlights

Text-to-Speech

Reading Statistics

Cross-Device Sync

Core Features

🤖 AI-Powered Reading

Intelligent Chat - Ask questions about your books, AI knows your position, selected text, and highlights
Semantic Search - Beyond keywords, vector retrieval + BM25 hybrid search
Instant Translation - AI translation or DeepL, 19 languages supported
Multiple AI Providers - OpenAI, Claude, Gemini, Ollama, DeepSeek
Skills System - Built-in skills (summarizer, concept explainer, character tracker, etc.) + create custom skills

📝 Annotation & Knowledge Management

5-Color Highlights - Yellow/Green/Blue/Pink/Purple, hover to preview notes
Markdown Notes - Rich text editor with toolbar, WYSIWYG
Multi-format Export - Markdown, HTML, JSON, Obsidian, Notion

🔊 Text-to-Speech (TTS)

Multiple Engines - Edge TTS, Browser TTS, DashScope (通义千问)
Voice Selection - 100+ voices in multiple languages
Speed Control - Adjustable playback speed
Background Playback - Listen while doing other things

📊 Reading Statistics

Reading Heatmap - Visualize your reading habits like GitHub contributions
Trend Charts - Track daily/weekly/monthly reading time
Streak Tracking - Longest consecutive reading days
Book Statistics - Time spent per book, completion rate

☁️ Cross-Device Sync

WebDAV Support - Sync your library, highlights, and notes across devices
Auto Sync - Automatic background synchronization
Conflict Resolution - Smart merge for concurrent edits

📚 Multi-Format Support

EPUB · PDF · MOBI · AZW · AZW3 · FB2 · FBZ · CBZ · TXT · UMD

TXT and UMD are imported by converting them to EPUB for reading, notes, search, and sync.

🎨 Customizable Experience

5 font themes (CJK optimized)
Light/Dark mode
Paginated/Continuous scroll
Keyboard shortcuts
English/Chinese interface

Quick Start

Download

Platform	Download
macOS (Apple Silicon)	Download .dmg
macOS (Intel)	Download .dmg
Windows	Download .msi
Linux	Download .AppImage
iOS	Join TestFlight
Android	Download .apk

Homebrew (macOS)

brew tap codedogQBY/readany
brew install --cask readany

3 Steps to Get Started

Import Books - Drag and drop files into library
Start Reading - Double-click to open, immersive experience
Configure AI (Optional) - Settings → AI → Enter API Key

Mobile Apps

ReadAny is now available on mobile devices!

Expo (React Native) Version:

# Clone and setup
git clone https://github.com/codedogQBY/ReadAny.git
cd ReadAny
pnpm install

# Install/run the development build on iOS
pnpm expo:ios

# Install/run the development build on iOS Simulator
pnpm expo:ios:simulator

# Install/run the development build on Android
# Start an Android emulator first, or connect a device.
pnpm expo:android

# Start Metro for the installed development build
pnpm expo:start

Mobile development uses an Expo development build with expo-dev-client, not Expo Go. Expo Go cannot load ReadAny's native modules and app configuration. Use pnpm expo:ios, pnpm expo:ios:simulator, or pnpm expo:android the first time, or whenever native dependencies/configuration change, then use pnpm expo:start for daily JS debugging in the installed ReadAny development app.

For simulators, use pnpm expo:ios:simulator on iOS. For Android, start an Android emulator first, then run pnpm expo:android.

Mobile app source lives in packages/app-expo.

AI Configuration

Provider	Get API Key
OpenAI	platform.openai.com
Anthropic Claude	console.anthropic.com
Google Gemini	aistudio.google.com
Ollama / DeepSeek	Local or custom endpoint

Tech Stack

Layer	Technology
Desktop	Tauri 2 (Rust)
Mobile	Expo (React Native) + Tauri Mobile
Frontend	React 19 + TypeScript
Build	Vite 7
Styling	Tailwind CSS 4 + Radix UI
State	Zustand
Database	SQLite
E-Book	foliate-js
AI/LLM	LangChain.js + LangGraph
Embeddings	Transformers.js

Development

# Clone
git clone https://github.com/codedogQBY/ReadAny.git
cd ReadAny

# Install
pnpm install

# Dev (Desktop)
pnpm tauri dev

# Dev (Mobile - Expo development build, not Expo Go)
# First install/run the native development build:
pnpm expo:ios
pnpm expo:ios:simulator
pnpm expo:android

# Then start Metro for the installed development app:
pnpm expo:start

# Build
pnpm tauri build

Mobile development uses expo-dev-client, so Expo Go is not supported. Re-run pnpm expo:ios, pnpm expo:ios:simulator, or pnpm expo:android after changing native dependencies, app.config.js, permissions, schemes, or build plugins.

Requirements: Node.js ≥18, pnpm ≥9, Rust (for Tauri), plus Xcode for iOS or Android Studio/SDK for Android mobile development.

Roadmap

Mobile apps — iOS and Android versions available
TTS (Text-to-Speech) — Edge TTS, multiple voices
Reading Statistics — Heatmap, trends, streaks
Skills System — Built-in + custom AI skills
WebDAV Sync — Cross-device synchronization
More AI models (Qwen, GLM, Llama)
PDF reflow/re-render
Plugin system
Cloud sync (official service)

Contributing

Contributions welcome! Bug reports, feature requests, pull requests all appreciated.

Fork → 2. Branch → 3. PR

Please run pnpm lint before submitting to ensure code style consistency.

License

This project is open source under the GNU General Public License v3.0. You are free to use, modify, and distribute the code, but any derivative works must also be open source under the same license.

Note: While the source code is freely available, the official app store versions may be offered for a fee to support ongoing development and cover certificate costs. You can always build the app yourself at no cost.

Acknowledgments

foliate-js - E-book rendering engine
Tauri - Cross-platform desktop framework
Expo - React Native development platform
LangChain.js - AI orchestration framework
Radix UI - Accessible UI components
Lucide - Icon library

Community

Thanks to linux.do — a vibrant Chinese tech community where you can learn about AI, development, and more.

☕ Support the Project

If you find ReadAny helpful, consider buying me a coffee to support ongoing development!

怎么上手

项目说明

核心能力

容易踩的坑

最近发版

新功能

优化

修复

English Release Notes

New Features

Improvements

Fixes

新功能

优化

修复

English Release Notes

New Features

Improvements

🚀 新功能

✨ 优化

🐛 修复

English Release Notes

🚀 New Features

✨ Improvements

🐛 Fixes

🚀 新功能

✨ 优化

🐛 修复

English Release Notes

🚀 New Features

✨ Improvements

🐛 Fixes

🚀 新功能

✨ 优化

🐛 修复

English Release Notes

🚀 New Features

✨ Improvements

🐛 Fixes

用户点评

README 原文

AI-Native Reading Workflow

Why ReadAny?

Comparison with Alternatives

Screenshots

Hero

AI-Powered Chat

Notes & Highlights

Text-to-Speech

Reading Statistics

Cross-Device Sync

Core Features

🤖 AI-Powered Reading

📝 Annotation & Knowledge Management

🔊 Text-to-Speech (TTS)

📊 Reading Statistics

☁️ Cross-Device Sync

📚 Multi-Format Support

🎨 Customizable Experience

Quick Start

Download

Homebrew (macOS)

3 Steps to Get Started

Mobile Apps

AI Configuration

Tech Stack

Development

Roadmap

Contributing

License

Acknowledgments

Community

☕ Support the Project

Star History

相关项目