rocky/OSGKeyboard
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: macOS architecture, cloud ASR/LLM providers, and 6-step iOS onboarding

Browse Source 
- Add macOS menu-bar dictation app with local ASR models (SenseVoice/Qwen3),
  global Option hotkey, and bottom overlay
- Add cloud ASR/LLM providers (Anthropic, Volcengine, Bailian, and more) with
  provider logos, model listing, and connection checks
- Add shared 7-day usage stats UI (UsageStatsCluster / SevenDayUsageChart)
- Add iOS onboarding step 6 for polish LLM setup; hide custom-language-model
  diagnostic toggle behind DEBUG
- Unify iOS onboarding tagline with the macOS brand line ("开口即文字。")
- Rewrite README (Chinese-first, product-oriented) and refresh GitHub Pages
This commit is contained in:
Rocky
2026-07-11 19:10:20 +08:00
parent cdf833935a
commit cc8dd1070a
116 changed files with 6659 additions and 2634 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/index.html
+174 -116
View File
@@ -1,13 +1,13 @@
<!DOCTYPE html>
<html lang="en">
<html lang="zh-Hans">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<!-- Primary SEO -->
<title>OSGKeyboard — On-device voice-to-text keyboard for iOS</title>
<meta name="description" content="OSGKeyboard is a source-available iOS keyboard that turns speech into polished text in any app. On-device recognition (iOS 26+), AI polish, translation, and a personal dictionary. On-device by default — your voice never leaves the device unless you opt into the cloud engine.">
<meta name="keywords" content="voice keyboard, voice to text, dictation, iOS keyboard, speech to text, on-device ASR, SpeechAnalyzer, AI polish, translation keyboard, open source keyboard, OSGKeyboard">
<title>OSGKeyboard — 开口即文字 · iOS &amp; Mac 语音输入</title>
<meta name="description" content="OSGKeyboard：iPhone、iPad、Mac 通用的语音输入工具。本地识别，AI 润色，任意 App 即说即写。Tap or hold Option — polished text lands at your cursor.">
<meta name="keywords" content="voice keyboard, dictation, speech to text, iOS keyboard, macOS dictation, 语音输入, 听写, OSGKeyboard, 开口即文字">
<meta name="author" content="OSGKeyboard">
<meta name="robots" content="index, follow, max-image-preview:large">
<meta name="theme-color" content="#0a0a0b">
@@ -18,8 +18,8 @@
<!-- Open Graph -->
<meta property="og:type" content="website">
<meta property="og:site_name" content="OSGKeyboard">
<meta property="og:title" content="OSGKeyboard — On-device voice-to-text keyboard for iOS">
<meta property="og:description" content="Turn speech into polished text in any app. On-device recognition (iOS 26+), AI polish, translation, personal dictionary. On-device by default — audio stays on your phone unless you opt into the cloud engine.">
<meta property="og:title" content="OSGKeyboard — 开口即文字 · iOS &amp; Mac 语音输入">
<meta property="og:description" content="本地识别，AI 润色，任意 App 即说即写。iPhone、iPad、Mac 通用。">
<meta property="og:url" content="https://hkgood.github.io/OSGKeyboard/">
<meta property="og:image" content="https://hkgood.github.io/OSGKeyboard/assets/app-icon.png">
<meta property="og:image:width" content="1024">
@@ -29,8 +29,8 @@
<!-- Twitter -->
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:title" content="OSGKeyboard — On-device voice-to-text keyboard for iOS">
<meta name="twitter:description" content="Turn speech into polished text in any app. On-device recognition, AI polish, translation, personal dictionary.">
<meta name="twitter:title" content="OSGKeyboard — 开口即文字">
<meta name="twitter:description" content="iPhone、iPad、Mac 语音输入。本地识别，AI 润色，任意 App 即说即写。">
<meta name="twitter:image" content="https://hkgood.github.io/OSGKeyboard/assets/app-icon.png">
<!-- Structured data -->
@@ -39,22 +39,22 @@
"@context": "https://schema.org",
"@type": "SoftwareApplication",
"name": "OSGKeyboard",
"operatingSystem": "iOS 26.0 or later",
"operatingSystem": "iOS 26.0 or later, macOS 14.0 or later",
"applicationCategory": "UtilitiesApplication",
"url": "https://hkgood.github.io/OSGKeyboard/",
"image": "https://hkgood.github.io/OSGKeyboard/assets/app-icon.png",
"description": "A source-available iOS keyboard that turns speech into polished text in any app. On-device recognition (iOS 26+), AI polish, translation, and a personal dictionary. Local or cloud engine.",
"softwareVersion": "0.3.6",
"description": "Voice input for iPhone, iPad, and Mac. On-device recognition, AI polish, personal dictionary, and iCloud sync. Speak in any app — polished text lands at your cursor.",
"softwareVersion": "0.5.3",
"isAccessibleForFree": true,
"offers": { "@type": "Offer", "price": "0", "priceCurrency": "USD" },
"author": { "@type": "Organization", "name": "OSGKeyboard" },
"featureList": [
"On-device speech recognition (SpeechAnalyzer, iOS 26+)",
"AI polish with three intensity levels",
"Post-polish translation",
"Personal dictionary",
"Long-lived Flow session",
"Local or cloud engine"
"Custom keyboard for iOS and iPadOS",
"macOS global dictation with Option hotkey",
"On-device speech recognition",
"AI text polish with multiple LLM providers",
"Personal dictionary with iCloud sync",
"Post-polish translation"
]
}
</script>
@@ -389,6 +389,30 @@
.step h3 { margin: 0 0 0.35rem; font-size: 1.05rem; font-weight: 650; }
.step p { margin: 0; color: var(--text-secondary); font-size: 0.95rem; }
/* Platforms */
.platforms {
display: grid;
grid-template-columns: repeat(2, 1fr);
gap: 1.25rem;
}
@media (max-width: 640px) { .platforms { grid-template-columns: 1fr; } }
.platform {
padding: 1.5rem;
border-radius: var(--radius);
border: 1px solid var(--line);
background: var(--bg-soft);
}
.platform-head {
display: flex;
align-items: center;
gap: 0.6rem;
margin-bottom: 0.75rem;
}
.platform-head .material-symbols-outlined { font-size: 22px; color: var(--accent); }
.platform h3 { margin: 0; font-size: 1.05rem; font-weight: 650; }
.platform p { margin: 0; font-size: 0.93rem; color: var(--text-secondary); line-height: 1.55; }
/* Closing */
.closing {
border-top: 1px solid var(--line);
@@ -424,8 +448,8 @@
</a>
<div class="nav-right">
<div class="lang-toggle" role="group" aria-label="Language">
<button type="button" class="active" data-lang="en">EN</button>
<button type="button" data-lang="zh">中文</button>
<button type="button" data-lang="en">EN</button>
<button type="button" class="active" data-lang="zh">中文</button>
</div>
<a class="ghost-link" href="https://github.com/hkgood/OSGKeyboard" rel="noopener">
<span class="material-symbols-outlined" aria-hidden="true">code</span>
@@ -442,24 +466,24 @@
<div>
<div class="eyebrow">
<span class="material-symbols-outlined" aria-hidden="true">verified</span>
<span data-i18n="hero.eyebrow">Source available · iOS 26+</span>
<span data-i18n="hero.eyebrow">iOS · iPadOS · macOS</span>
</div>
<h1 data-i18n="hero.title">Speak naturally.<br>Get polished text anywhere.</h1>
<p class="lead" data-i18n="hero.lead">A source-available keyboard that transcribes your voice on-device, then polishes — and optionally translates — the result. Works in every app you type in.</p>
<h1 data-i18n="hero.title">开口即文字。</h1>
<p class="lead" data-i18n="hero.lead">iPhone、iPad、Mac 通用的语音输入工具。本地识别，AI 润色，任意 App 即说即写。</p>
<div class="hero-cta">
<a class="btn btn-primary" href="https://github.com/hkgood/OSGKeyboard" rel="noopener">
<span class="material-symbols-outlined" aria-hidden="true">download</span>
<span data-i18n="hero.cta.primary">Get it on GitHub</span>
<span data-i18n="hero.cta.primary">立即获取</span>
</a>
<a class="btn btn-quiet" href="privacy/">
<span class="material-symbols-outlined" aria-hidden="true">shield</span>
<span data-i18n="hero.cta.secondary">Privacy</span>
<span data-i18n="hero.cta.secondary">隐私政策</span>
</a>
</div>
<div class="hero-meta">
<span><span class="material-symbols-outlined" aria-hidden="true">devices</span><span data-i18n="hero.meta1">On-device recognition</span></span>
<span><span class="material-symbols-outlined" aria-hidden="true">bolt</span><span data-i18n="hero.meta2">Local or cloud AI</span></span>
<span><span class="material-symbols-outlined" aria-hidden="true">lock</span><span data-i18n="hero.meta3">On-device by default — audio stays on your phone</span></span>
<span><span class="material-symbols-outlined" aria-hidden="true">devices</span><span data-i18n="hero.meta1">默认本地识别</span></span>
<span><span class="material-symbols-outlined" aria-hidden="true">auto_awesome</span><span data-i18n="hero.meta2">AI 自动润色</span></span>
<span><span class="material-symbols-outlined" aria-hidden="true">apps</span><span data-i18n="hero.meta3">任意 App 可用</span></span>
</div>
</div>
@@ -483,42 +507,66 @@
</div>
</section>
<!-- Platforms -->
<section class="block wrap">
<div class="block-head">
<h2 data-i18n="platforms.title">两个平台，同一种体验</h2>
<p data-i18n="platforms.subtitle">手机上用键盘，Mac 上按住 Option — 说完即落字。</p>
</div>
<div class="platforms">
<div class="platform">
<div class="platform-head">
<span class="material-symbols-outlined" aria-hidden="true">smartphone</span>
<h3 data-i18n="platform.ios.t">iOS / iPadOS</h3>
</div>
<p data-i18n="platform.ios.b">自定义键盘 + Flow 会话。点按麦克风说话，连续听写不用跳 App。灵动岛显示录音状态。</p>
</div>
<div class="platform">
<div class="platform-head">
<span class="material-symbols-outlined" aria-hidden="true">laptop_mac</span>
<h3 data-i18n="platform.mac.t">macOS</h3>
</div>
<p data-i18n="platform.mac.b">菜单栏常驻，按住 Option 全局听写。底部浮层实时反馈，支持 SenseVoice / Qwen3 本地模型。</p>
</div>
</div>
</section>
<!-- Features -->
<section class="block wrap">
<div class="block-head">
<h2 data-i18n="features.title">Everything you need to type by voice</h2>
<p data-i18n="features.subtitle">On-device by default, refined by AI when you want it — in a single, native keyboard.</p>
<h2 data-i18n="features.title">开口即写，所需尽在此</h2>
<p data-i18n="features.subtitle">本地优先，AI 加持 — 从识别到润色，一条链路搞定。</p>
</div>
<div class="features">
<div class="feature">
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">mic</span></div>
<h3 data-i18n="feat.dictation.t">Tap-to-talk dictation</h3>
<p data-i18n="feat.dictation.b">Tap to start, tap to stop. Up to 3.5 minutes per take, with a live countdown.</p>
<h3 data-i18n="feat.dictation.t">点按听写</h3>
<p data-i18n="feat.dictation.b">点一下开始，再点一下结束。单次最长 3.5 分钟，倒计时实时显示。</p>
</div>
<div class="feature">
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">devices</span></div>
<h3 data-i18n="feat.ondevice.t">On-device recognition</h3>
<p data-i18n="feat.ondevice.b">iOS 26 SpeechAnalyzer transcribes on your device by default — audio is uploaded only if you explicitly enable the cloud engine.</p>
<h3 data-i18n="feat.ondevice.t">本地识别</h3>
<p data-i18n="feat.ondevice.b">iOS 端侧 SpeechAnalyzer；Mac 可选 SenseVoice / Qwen3。默认不上传录音。</p>
</div>
<div class="feature">
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">auto_awesome</span></div>
<h3 data-i18n="feat.polish.t">AI polish</h3>
<p data-i18n="feat.polish.b">Punctuation, structure, and clarity in three intensity levels. Polish sends only text — on the default engine your audio never goes online.</p>
<h3 data-i18n="feat.polish.t">AI 润色</h3>
<p data-i18n="feat.polish.b">自动补标点、整理结构。内置润色开箱即用，也可接入 DeepSeek、OpenAI 等 API。</p>
</div>
<div class="feature">
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">translate</span></div>
<h3 data-i18n="feat.translate.t">Built-in translation</h3>
<p data-i18n="feat.translate.b">Translate after polish into English, 中文, 日本語, 한국어 and more — right from the keyboard.</p>
<h3 data-i18n="feat.translate.t">润色后翻译</h3>
<p data-i18n="feat.translate.b">润色完成后一键翻译为 English、中文、日本語、한국어 等。</p>
</div>
<div class="feature">
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">menu_book</span></div>
<h3 data-i18n="feat.dict.t">Personal dictionary</h3>
<p data-i18n="feat.dict.b">Keep names, products, and jargon accurate. The AI preserves your terms every time.</p>
<h3 data-i18n="feat.dict.t">个性词库</h3>
<p data-i18n="feat.dict.b">人名、产品名、专业术语 — AI 润色时原样保留，iCloud 多设备同步。</p>
</div>
<div class="feature">
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">all_inclusive</span></div>
<h3 data-i18n="feat.flow.t">Flow session</h3>
<p data-i18n="feat.flow.b">A long-lived session keeps the recorder warm, so you can dictate take after take.</p>
<div class="ficon"><span class="material-symbols-outlined" aria-hidden="true">cloud_sync</span></div>
<h3 data-i18n="feat.sync.t">iCloud 同步</h3>
<p data-i18n="feat.sync.b">设置、词库、听写历史在多设备间自动同步，换机无缝衔接。</p>
</div>
</div>
</section>
@@ -526,21 +574,21 @@
<!-- How it works -->
<section class="block wrap">
<div class="block-head">
<h2 data-i18n="steps.title">Three steps to your first message</h2>
<p data-i18n="steps.subtitle">Set up once, then dictate anywhere you type.</p>
<h2 data-i18n="steps.title">三步，开始第一句话</h2>
<p data-i18n="steps.subtitle">设置一次，之后在任意输入处开口即写。</p>
</div>
<div class="steps">
<div class="step">
<h3 data-i18n="step1.t">Enable the keyboard</h3>
<p data-i18n="step1.b">Add OSGKeyboard in iOS Settings and allow Full Access — needed for the mic and network.</p>
<h3 data-i18n="step1.t">安装并授权</h3>
<p data-i18n="step1.b">iOS 添加键盘并开启完全访问；Mac 授予麦克风与辅助功能权限。</p>
</div>
<div class="step">
<h3 data-i18n="step2.t">Choose your engine</h3>
<p data-i18n="step2.b">Local: on-device ASR with built-in polish, no API key. Cloud: use your own OpenAI-compatible key.</p>
<h3 data-i18n="step2.t">选引擎与润色</h3>
<p data-i18n="step2.b">本地识别 + 内置润色零配置可用；也可填入自己的 API Key 接入任意兼容模型。</p>
</div>
<div class="step">
<h3 data-i18n="step3.t">Tap to talk</h3>
<p data-i18n="step3.b">Switch to OSGKeyboard, tap the mic, speak, tap again. Polished text lands at your cursor.</p>
<h3 data-i18n="step3.t">开口说话</h3>
<p data-i18n="step3.b">切换到 OSGKeyboard 键盘，或按住 Option 键。润色好的文字自动插入光标处。</p>
</div>
</div>
</section>
@@ -549,13 +597,13 @@
<section class="closing">
<div class="wrap closing-inner">
<div>
<span class="lock"><span class="material-symbols-outlined" aria-hidden="true">lock</span><span data-i18n="close.badge">Privacy by design</span></span>
<h2 data-i18n="close.title">Your voice stays yours</h2>
<p data-i18n="close.body">Speech is processed on-device by default (the optional cloud engine uploads recordings to the provider you configure). We never log ordinary keystrokes; polish and translation receive only transcribed text.</p>
<span class="lock"><span class="material-symbols-outlined" aria-hidden="true">lock</span><span data-i18n="close.badge">隐私优先</span></span>
<h2 data-i18n="close.title">你的声音，始终属于你</h2>
<p data-i18n="close.body">默认在设备端识别，录音不上传。润色只发送转写文字，不记录日常击键。云端引擎需你主动开启，并适用所配置服务商的隐私条款。</p>
</div>
<a class="btn btn-quiet" href="privacy/">
<span class="material-symbols-outlined" aria-hidden="true">description</span>
<span data-i18n="close.cta">Read the privacy policy</span>
<span data-i18n="close.cta">查看隐私政策</span>
</a>
</div>
</section>
@@ -563,7 +611,7 @@
<footer class="wrap">
<div class="footer-inner">
<span data-i18n="footer.copy">© OSGKeyboard · v0.3.6 · source available, non-commercial</span>
<span data-i18n="footer.copy">© OSGKeyboard · v0.5.3 · 源码可见</span>
<span class="footer-links">
<a href="https://github.com/hkgood/OSGKeyboard" rel="noopener" data-i18n="footer.source">Source</a>
<a href="privacy/" data-i18n="footer.privacy">Privacy</a>
@@ -576,90 +624,102 @@
const copy = {
en: {
"nav.github": "GitHub",
"hero.eyebrow": "Source available · iOS 26+",
"hero.title": "Speak naturally.<br>Get polished text anywhere.",
"hero.lead": "A source-available keyboard that transcribes your voice on-device, then polishes — and optionally translates — the result. Works in every app you type in.",
"hero.cta.primary": "Get it on GitHub",
"hero.eyebrow": "iOS · iPadOS · macOS",
"hero.title": "Speak it. It's typed.",
"hero.lead": "Voice input for iPhone, iPad, and Mac. On-device recognition, AI polish — speak in any app.",
"hero.cta.primary": "Get it now",
"hero.cta.secondary": "Privacy",
"hero.meta1": "On-device recognition",
"hero.meta2": "Local or cloud AI",
"hero.meta3": "On-device by default — audio stays on your phone",
"hero.meta1": "On-device by default",
"hero.meta2": "AI polish",
"hero.meta3": "Works in every app",
"mock.app": "Notes",
"mock.line": "Tomorrow at 3 PM, sync with the design team on the new onboarding flow.",
"mock.chipEngine": "On-device",
"mock.chipEngine": "Local",
"mock.chipTranslate": "Translate",
"platforms.title": "Two platforms, one experience",
"platforms.subtitle": "Keyboard on phone, hold Option on Mac — speak and it's typed.",
"platform.ios.t": "iOS / iPadOS",
"platform.ios.b": "Custom keyboard + Flow session. Tap the mic, dictate take after take without leaving your app. Live Activity on Dynamic Island.",
"platform.mac.t": "macOS",
"platform.mac.b": "Menu-bar app with global Option hotkey. Bottom overlay with live feedback. Optional SenseVoice / Qwen3 local models.",
"features.title": "Everything you need to type by voice",
"features.subtitle": "On-device by default, refined by AI when you want it — in a single, native keyboard.",
"feat.dictation.t": "Tap-to-talk dictation",
"feat.dictation.b": "Tap to start, tap to stop. Up to 3.5 minutes per take, with a live countdown.",
"features.subtitle": "Local first, AI refined — recognition to polish in one flow.",
"feat.dictation.t": "Tap-to-talk",
"feat.dictation.b": "Tap to start, tap to stop. Up to 3.5 minutes per take with a live countdown.",
"feat.ondevice.t": "On-device recognition",
"feat.ondevice.b": "iOS 26 SpeechAnalyzer transcribes on your device by default — audio is uploaded only if you explicitly enable the cloud engine.",
"feat.ondevice.b": "iOS SpeechAnalyzer; Mac SenseVoice / Qwen3 optional. Audio stays local by default.",
"feat.polish.t": "AI polish",
"feat.polish.b": "Punctuation, structure, and clarity in three intensity levels. Polish sends only text — on the default engine your audio never goes online.",
"feat.translate.t": "Built-in translation",
"feat.translate.b": "Translate after polish into English, 中文, 日本語, 한국어 and more — right from the keyboard.",
"feat.polish.b": "Punctuation and structure, three intensity levels. Built-in polish or bring your own DeepSeek / OpenAI API.",
"feat.translate.t": "Post-polish translation",
"feat.translate.b": "Translate into English, 中文, 日本語, 한국어 and more after polish.",
"feat.dict.t": "Personal dictionary",
"feat.dict.b": "Keep names, products, and jargon accurate. The AI preserves your terms every time.",
"feat.flow.t": "Flow session",
"feat.flow.b": "A long-lived session keeps the recorder warm, so you can dictate take after take.",
"feat.dict.b": "Names, products, jargon — preserved verbatim by the AI, synced via iCloud.",
"feat.sync.t": "iCloud sync",
"feat.sync.b": "Settings, dictionary, and history sync across devices seamlessly.",
"steps.title": "Three steps to your first message",
"steps.subtitle": "Set up once, then dictate anywhere you type.",
"step1.t": "Enable the keyboard",
"step1.b": "Add OSGKeyboard in iOS Settings and allow Full Access — needed for the mic and network.",
"step2.t": "Choose your engine",
"step2.b": "Local: on-device ASR with built-in polish, no API key. Cloud: use your own OpenAI-compatible key.",
"step3.t": "Tap to talk",
"step3.b": "Switch to OSGKeyboard, tap the mic, speak, tap again. Polished text lands at your cursor.",
"close.badge": "Privacy by design",
"steps.subtitle": "Set up once, then dictate anywhere.",
"step1.t": "Install & authorize",
"step1.b": "Add the iOS keyboard with Full Access; grant mic + Accessibility on Mac.",
"step2.t": "Pick engine & polish",
"step2.b": "Local ASR + built-in polish works out of the box. Or add your own API key.",
"step3.t": "Start talking",
"step3.b": "Switch to OSGKeyboard, or hold Option on Mac. Polished text lands at your cursor.",
"close.badge": "Privacy first",
"close.title": "Your voice stays yours",
"close.body": "Speech is processed on-device by default (the optional cloud engine uploads recordings to the provider you configure). We never log ordinary keystrokes; polish and translation receive only transcribed text.",
"close.body": "On-device recognition by default — audio is not uploaded. Polish receives text only; we never log keystrokes. Cloud engine requires explicit opt-in.",
"close.cta": "Read the privacy policy",
"footer.copy": "© OSGKeyboard · v0.3.6 · source available, non-commercial",
"footer.copy": "© OSGKeyboard · v0.5.3 · source available",
"footer.source": "Source",
"footer.privacy": "Privacy",
"footer.license": "License"
},
zh: {
"nav.github": "GitHub",
"hero.eyebrow": "源码可见 · iOS 26+",
"hero.title": "自然地说，<br>随处得到润色好的文字。",
"hero.lead": "源码可见的语音键盘：在设备端转写你的语音，再润色，并可选翻译。适用于所有需要输入的 App。",
"hero.cta.primary": "前往 GitHub",
"hero.eyebrow": "iOS · iPadOS · macOS",
"hero.title": "开口即文字。",
"hero.lead": "iPhone、iPad、Mac 通用的语音输入工具。本地识别，AI 润色，任意 App 即说即写。",
"hero.cta.primary": "立即获取",
"hero.cta.secondary": "隐私政策",
"hero.meta1": "设备端识别",
"hero.meta2": "本地或云端 AI",
"hero.meta3": "默认端侧识别，音频留在手机",
"hero.meta1": "默认本地识别",
"hero.meta2": "AI 自动润色",
"hero.meta3": "任意 App 可用",
"mock.app": "备忘录",
"mock.line": "明天下午三点，和设计团队同步新引导流程。",
"mock.chipEngine": "本地",
"mock.chipTranslate": "翻译",
"features.title": "语音输入所需的一切",
"features.subtitle": "默认全程端侧，需要时再交给 AI 润色——都在一个原生键盘里完成。",
"platforms.title": "两个平台，同一种体验",
"platforms.subtitle": "手机上用键盘，Mac 上按住 Option — 说完即落字。",
"platform.ios.t": "iOS / iPadOS",
"platform.ios.b": "自定义键盘 + Flow 会话。点按麦克风说话，连续听写不用跳 App。灵动岛显示录音状态。",
"platform.mac.t": "macOS",
"platform.mac.b": "菜单栏常驻，按住 Option 全局听写。底部浮层实时反馈，支持 SenseVoice / Qwen3 本地模型。",
"features.title": "开口即写，所需尽在此",
"features.subtitle": "本地优先，AI 加持 — 从识别到润色，一条链路搞定。",
"feat.dictation.t": "点按听写",
"feat.dictation.b": "点一下开始，再点一下结束。单次最长 3.5 分钟，倒计时实时显示。",
"feat.ondevice.t": "设备端识别",
"feat.ondevice.b": "默认由 iOS 26 SpeechAnalyzer 在本机转写——仅当你显式开启云端引擎时，录音才会上传到你配置的服务商。",
"feat.ondevice.t": "本地识别",
"feat.ondevice.b": "iOS 端侧 SpeechAnalyzer；Mac 可选 SenseVoice / Qwen3。默认不上传录音。",
"feat.polish.t": "AI 润色",
"feat.polish.b": "自动补全标点、结构与表达，分三档强度。润色仅发送文字；默认引擎下音频不联网。",
"feat.translate.t": "内置翻译",
"feat.translate.b": "润色后可翻译为 English、中文、日本語、한국어 等——直接在键盘上完成。",
"feat.polish.b": "自动补标点、整理结构。内置润色开箱即用，也可接入 DeepSeek、OpenAI 等 API。",
"feat.translate.t": "润色后翻译",
"feat.translate.b": "润色完成后一键翻译为 English、中文、日本語、한국어 等。",
"feat.dict.t": "个性词库",
"feat.dict.b": "让人名、产品名与专有名词保持准确，AI 每次都会原样保留你的词汇。",
"feat.flow.t": "Flow 会话",
"feat.flow.b": "长生命周期会话让录音随时就绪，连续多次听写无需跳回主 App。",
"steps.title": "三步发出第一条消息",
"steps.subtitle": "只需设置一次，随后在任意输入处开口即写。",
"step1.t": "启用键盘",
"step1.b": "在 iOS 设置中添加 OSGKeyboard 并允许完全访问——麦克风与网络所需。",
"step2.t": "选择引擎",
"step2.b": "本地：设备端 ASR + 内置润色，无需 API Key。云端：使用你自己的 OpenAI 兼容 Key。",
"step3.t": "点按说话",
"step3.b": "切换到 OSGKeyboard，点麦克风，说话，再点结束，润色好的文字自动插入光标处。",
"close.badge": "隐私优先设计",
"feat.dict.b": "人名、产品名、专业术语 — AI 润色时原样保留，iCloud 多设备同步。",
"feat.sync.t": "iCloud 同步",
"feat.sync.b": "设置、词库、听写历史在多设备间自动同步，换机无缝衔接。",
"steps.title": "三步，开始第一句话",
"steps.subtitle": "设置一次，之后在任意输入处开口即写。",
"step1.t": "安装并授权",
"step1.b": "iOS 添加键盘并开启完全访问；Mac 授予麦克风与辅助功能权限。",
"step2.t": "选引擎与润色",
"step2.b": "本地识别 + 内置润色零配置可用；也可填入自己的 API Key 接入任意兼容模型。",
"step3.t": "开口说话",
"step3.b": "切换到 OSGKeyboard 键盘，或按住 Option 键。润色好的文字自动插入光标处。",
"close.badge": "隐私优先",
"close.title": "你的声音，始终属于你",
"close.body": "默认在设备端处理语音（可选的云端引擎会把录音上传到你配置的服务商）。我们不记录普通击键；润色与翻译只接收转写文字。",
"close.body": "默认在设备端识别，录音不上传。润色只发送转写文字，不记录日常击键。云端引擎需你主动开启，并适用所配置服务商的隐私条款。",
"close.cta": "查看隐私政策",
"footer.copy": "© OSGKeyboard · v0.3.6 · 源码可见，禁止商用",
"footer.copy": "© OSGKeyboard · v0.5.3 · 源码可见",
"footer.source": "源代码",
"footer.privacy": "隐私",
"footer.license": "许可协议"
@@ -667,13 +727,12 @@
};
function applyLang(lang) {
const strings = copy[lang] || copy.en;
const strings = copy[lang] || copy.zh;
document.documentElement.lang = lang === "zh" ? "zh-Hans" : "en";
document.querySelectorAll("[data-i18n]").forEach((el) => {
const value = strings[el.getAttribute("data-i18n")];
if (value == null) return;
if (el.getAttribute("data-i18n") === "hero.title") el.innerHTML = value;
else el.textContent = value;
el.textContent = value;
});
document.querySelectorAll(".lang-toggle button").forEach((btn) => {
btn.classList.toggle("active", btn.dataset.lang === lang);
@@ -686,8 +745,7 @@
});
const saved = (() => { try { return localStorage.getItem("osg-lang"); } catch (_) { return null; } })();
const zh = navigator.language && navigator.language.toLowerCase().startsWith("zh");
applyLang(saved || (zh ? "zh" : "en"));
applyLang(saved || "zh");
</script>
</body>
</html>
docs/local-asr-architecture.md
+7 -7
View File
@@ -101,7 +101,7 @@ flowchart LR
关键代码：
- [`OSGKeyboardMac/MacDictationPipeline.swift`](../OSGKeyboardMac/MacDictationPipeline.swift) — 本地分支未传 `personalDictionary`
- [`OSGKeyboardMac/MacLocalASRService.swift`](../OSGKeyboardMac/MacLocalASRService.swift) — Qwen3 / Apple Speech 二选一
- [`OSGKeyboardMac/MacLocalASRService.swift`](../OSGKeyboardMac/MacLocalASRService.swift) — Sherpa catalog models / Apple Speech fallback
- [`OSGKeyboardShared/Models/PersonalDictionary+ASRBias.swift`](../OSGKeyboardShared/Models/PersonalDictionary+ASRBias.swift) — 云侧 `asrHotwords` / `asrPromptBias` / 阿里热词表
- [`OSGKeyboardShared/Services/PolishingService.swift`](../OSGKeyboardShared/Services/PolishingService.swift) — 润色层消费 `dictionary.promptFragment()`
@@ -113,9 +113,9 @@ flowchart LR
### 3.4 本地模型现状
- 默认路径：`~/Library/Application Support/OSGKeyboard/models/qwen3-asr-1.7b-mlx`
- 校验：`config.json`、`model.safetensors`、`vocab.json`、`merges.txt`
- Qwen3 引擎：[`MacQwen3LocalASR.swift`](../OSGKeyboardMac/MacQwen3LocalASR.swift) / [`MacQwen3ASREngine.swift`](../OSGKeyboardMac/MacQwen3ASREngine.swift) — **无 prompt / hotword 入参**
- 默认模型：`sherpa-qwen3-0.6b-int8`（catalog 管理，安装于 `~/Library/Application Support/OSGKeyboard/`）
- 运行时：`sherpa-onnx-offline` 二进制（按架构下载，见 `local-asr-catalog.json`）
- Sherpa 引擎：[`MacSherpaLocalASR.swift`](../OSGKeyboardMac/MacSherpaLocalASR.swift) / [`MacSherpaONNXRunner.swift`](../OSGKeyboardMac/MacSherpaONNXRunner.swift) — Qwen3 支持 recognizer-scoped hotwords
---
@@ -369,8 +369,8 @@ OSG 已有 `PersonalDictionary.Entry.Source.recentEdit` 与合并逻辑，可与
~/Library/Application Support/OSGKeyboard/
LocalASRModels/
manifest.json # 已安装模型、版本、backend、capabilities
qwen3-mlx-1.7b/ # 当前 MLX 布局（可与现路径兼容）
sherpa-qwen3-0.6b/
runtimes/sherpa-onnx-1.13.4-macos-arm64/
models/sherpa-qwen3-0.6b-int8/
sherpa-sensevoice-small/
```
@@ -567,7 +567,7 @@ flowchart TD
|------|------|
| Mac 听写管道 | `OSGKeyboardMac/MacDictationPipeline.swift` |
| 本地 ASR 入口 | `OSGKeyboardMac/MacLocalASRService.swift` |
| Qwen3 MLX | `OSGKeyboardMac/MacQwen3LocalASR.swift`, `MacQwen3ASREngine.swift` |
| Sherpa local ASR | `OSGKeyboardMac/MacSherpaLocalASR.swift`, `MacSherpaONNXRunner.swift` |
| Apple Speech fallback | `OSGKeyboardMac/MacSpeechLocalASR.swift` |
| 用户词库 | `OSGKeyboardShared/Models/PersonalDictionary.swift` |
| 云 bias | `OSGKeyboardShared/Models/PersonalDictionary+ASRBias.swift` |