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 desktop PDF controls, zoom, paginated jumps, and EPUB image resource handling
- Vectorization and knowledge retrieval are more robust with better chapter extraction, remote embedding requests, PDF/EPUB chapter fallback, and large-file handling
- Sync and stats refresh are more reliable around WebDAV directories, synced local paths, and reading stats after book metadata changes
Fixes
- Fixed in-progress AI Q&A disappearing after leaving and reopening the chat / reader screen
- Fixed cloud TTS pause, cancellation, early sentence advance, and progress sync stability issues
- Fixed stale audio or duplicate playback after switching TTS engine, voice, rate, or pitch during playback
- Fixed mobile TTS state resync after returning from the background
- Fixed compatibility issues in mobile EPUB vertical layout, fixed layout, image-only pages, RTL page turns, and CFI jumps
- Fixed selected desktop PDF controls, page scrolling, zoom, and reading-location persistence issues
- Fixed selected EPUB image resource, script handling, and first-pagination settings application issues
- Fixed WebDAV remote root casing, repeated MKCOL handling, and synced local file paths
- Fixed Microsoft Chinese translation target mapping
完整变更对比 / Full changelog:
v1.3.4...v1.3.5
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.