AI 键盘模式

• 新增临时多轮 AI 输入面：把口述问题直接交给已配置模型，滚动查看最新答案，再通过「插入」或「发送」写入当前输入框
• 只有真正插入的答案会进入历史与 AI 字数统计
• 支持服务商侧联网搜索（在可用时自动启用），并强制开启 thinking；失败时会静默改为无搜索重试

API Key 与模型

• 润色与 AI 模式均需自行填写 API Key；未配置时首页与麦克风上方会提示，听写仍可插入原始识别结果
• 更新多家服务商的默认模型（含 OpenAI gpt-5.4-mini）；已保存的模型名不受影响

iPad 体验

• 语音面与打字面新增系统 🌐 键（轻点切换键盘，长按打开系统键盘列表）
• 打字布局适配 iPad：更高键行、横竖屏缩进、首行数字角标，并铺满宿主宽度
• 打字顶栏新增撤销 / 重做 / 拷贝 / 剪切

编辑上次输入

• 长按麦克风口述如何修改最近一次已验证输入，预览后可替换或追加，且不暴露无关输入框内容

AI keyboard mode

• Temporary multi-turn AI surface: send spoken questions to your configured model, review the latest answer, then Insert or Send into the host field
• Only inserted answers enter history and AI character statistics
• Provider-side web search when available, with thinking forced on; silent retry without search on failure

API keys & models

• Polish and AI mode require your own API key; Home and the mic line tip you when it’s missing, while dictation still inserts raw ASR
• Refreshed default models across providers (including OpenAI gpt-5.4-mini); saved model names are unchanged

iPad

• System 🌐 key on voice and typing surfaces (tap to switch, long-press for the system picker)
• Typing layout adapted for iPad: taller rows, orientation-aware inset, number overlays, full host width
• Undo / redo / copy / cut cluster on the typing top bar

Edit last input

• Long-press the mic to describe changes to the last verified insertion, then preview and replace or append — without exposing unrelated field text

全新保活流程

• 键盘保活流程全新升级，大幅提升后台保活逻辑，大幅降低耗电量并提升稳定性

新增键盘输入功能

• 新增拼音与英文键盘：支持全拼、微软/搜狗双拼、词组候选、可选模糊音；拼音与英文均支持自定义词库
• 键盘布局整体升级，语音 / 中文 / 英文外观统一；左侧 Logo 可跳转主 App，右侧改为输入方式切换与翻译入口
• 新增中文候选词面板，候选池更大，长词和更多备选更容易找到；多字输入时可分词选择
• 新增英文联想补全、自动纠错与候选撤销，输入更快更准
• 新增键盘防误触和误输入纠正，整体体验更接近系统键盘
• 新增按键音效与震动，以及轻 / 强震动设置
• 新增长按删除加速，改长文更高效
• 新增英文键盘智能大写：连续大写、句首自动大写
• 新增撤销上次输入按钮：麦克风旁可一键撤回最近一次听写或编辑结果
• 新增键盘偏好设置，支持记住上次使用的语音或文字界面，再次打开无需反复切换

编辑上次输入

• 刚用 OSGKeyboard 输入过内容时，麦克风上方会提示「长按可编辑上一条」
• 长按麦克风进入编辑，说出想怎么改（如「改得更正式些」「结尾加一句问候」）；短按仍是普通听写
• 说完再点一下结束录音，稍候即可左右滑动对比原文 / 编辑后
• 确认后点应用编辑替换原文，或选择插入当前位置；右上角关闭可放弃本次编辑

润色模式整体升级

• 大幅升级趣味润色（含直男癌等），风格更鲜明，同时减少事实失真
• 润色强度简化为轻度 / 重度：轻度优先保真，重度释放五种趣味风格的完整创意
• 新增同音 / 近音纠错，语音识别后的错字修复更准
• 长文本润色与翻译分段更清晰，列表结构更易读，翻译后保留段落格式

应用内更新说明

• 版本更新后自动展示更新说明；设置页版本行也可随时查看，跟随应用内语言与深浅色

设置优化

• 设置页层级与卡片布局更清晰，转写与润色入口更好找
• 简化首次设置：未完成语音配置时仍可正常打字
• 文本输入设置页的方案、模糊音与资源文案跟随应用内语言

Redesigned keep-alive flow

• Keyboard keep-alive is rebuilt for more reliable background continuity, lower power use, and better stability

Typing keyboards

• New Pinyin and English keyboards: full Pinyin, Microsoft/Sogou double Pinyin, phrase candidates, optional fuzzy Pinyin; both support custom dictionaries
• Unified keyboard chrome across Voice / Chinese / English; left logo opens the main app; right side switches input mode and opens translation
• New Chinese candidate panel with a larger pool for long phrases and alternatives; multi-character input supports segmented selection
• English autocomplete, autocorrect, and candidate undo for faster, more accurate typing
• Anti-mis-tap and mistype correction for a closer-to-system keyboard feel
• New key sounds and haptics, with light / strong haptic options
• Accelerating long-press delete for faster long-text edits
• Smarter English capitalization: caps lock and sentence-start autocapitalization
• New undo last input button beside the mic to roll back the latest dictation or edit result
• New Keyboard Preferences: remember the last Voice or Typing surface so you don’t keep switching

Edit last input

• Right after you insert text with OSGKeyboard, the mic shows “Hold to edit your last input”
• Long-press the mic to start editing and say how to change it (e.g. “make it more formal”, “add a greeting at the end”); a short press is still plain dictation
• Tap once more to finish, then swipe to compare Original / Edited
• Confirm to apply the edit (replace), or choose insert at cursor; close from the top-right to discard

Polish upgrades

• Stronger fun polish styles (including Straight-Talk Rescue / 直男癌), with clearer voice and less factual drift
• Intensity simplified to Light / Heavy: Light prioritizes fidelity; Heavy unlocks the full creativity of the five fun styles
• New homophone / near-homophone correction after speech recognition
• Clearer segmentation for long polish and translation; lists read better; translations keep paragraph structure

In-app release notes

• Release notes appear automatically after an update, and the Settings version row opens them any time, following your in-app language and theme

Settings

• Clearer Settings hierarchy and cards; transcription and polish are easier to find
• Simpler first-run setup: you can type even before voice setup is finished
• Schema, fuzzy-Pinyin, and resource strings on the Text Input settings page follow the in-app language