diff --git a/addon/content/chatpapers.css b/addon/content/chatpapers.css
index 57e0b58..7b830c9 100644
--- a/addon/content/chatpapers.css
+++ b/addon/content/chatpapers.css
@@ -752,3 +752,215 @@
.chatpapers-cite-btn {
flex-shrink: 0;
}
+
+/* --- Voice lecture pane --- */
+.chatpapers-lecture-root {
+ gap: 10px;
+}
+
+.chatpapers-lecture-empty {
+ padding: 20px 12px;
+}
+
+.chatpapers-lecture-status {
+ padding: 8px 10px;
+ border-radius: 8px;
+ font-size: 12px;
+ line-height: 1.45;
+ background: color-mix(in srgb, #1f7a8c 10%, transparent);
+ border: 1px solid color-mix(in srgb, #1f7a8c 22%, transparent);
+}
+
+.chatpapers-lecture-status[data-status="failed"] {
+ background: color-mix(in srgb, #dc2626 10%, transparent);
+ border-color: color-mix(in srgb, #dc2626 25%, transparent);
+}
+
+.chatpapers-lecture-status[data-status="ready"] {
+ background: color-mix(in srgb, #16a34a 10%, transparent);
+ border-color: color-mix(in srgb, #16a34a 25%, transparent);
+}
+
+.chatpapers-lecture-actions {
+ display: flex;
+ gap: 8px;
+}
+
+.chatpapers-lecture-primary {
+ display: inline-flex;
+ align-items: center;
+ gap: 6px;
+ padding: 8px 14px;
+ border: none;
+ border-radius: 8px;
+ background: #1f7a8c;
+ color: #fff;
+ font-size: 13px;
+ font-weight: 600;
+ cursor: pointer;
+}
+
+.chatpapers-lecture-primary:hover:not(:disabled) {
+ background: #186678;
+}
+
+.chatpapers-lecture-status-text {
+ white-space: pre-line;
+}
+
+.chatpapers-lecture-primary:disabled {
+ opacity: 0.55;
+ cursor: not-allowed;
+}
+
+.chatpapers-lecture-secondary {
+ display: inline-flex;
+ align-items: center;
+ gap: 6px;
+ padding: 8px 14px;
+ border: 1px solid color-mix(in srgb, #1f7a8c 35%, transparent);
+ border-radius: 8px;
+ background: transparent;
+ color: inherit;
+ font-size: 13px;
+ font-weight: 600;
+ cursor: pointer;
+}
+
+.chatpapers-lecture-secondary:hover:not(:disabled) {
+ background: color-mix(in srgb, #1f7a8c 10%, transparent);
+}
+
+.chatpapers-lecture-secondary:disabled {
+ opacity: 0.55;
+ cursor: not-allowed;
+}
+
+.chatpapers-lecture-tts-meta {
+ font-size: 12px;
+ opacity: 0.75;
+}
+
+.chatpapers-lecture-audio {
+ width: 100%;
+ margin-top: 4px;
+}
+
+.chatpapers-lecture-btn-icon {
+ flex-shrink: 0;
+}
+
+.chatpapers-lecture-hint {
+ font-size: 11px;
+ opacity: 0.65;
+ line-height: 1.4;
+}
+
+.chatpapers-lecture-player {
+ display: flex;
+ flex-direction: column;
+ gap: 8px;
+ margin-top: 4px;
+}
+
+.chatpapers-lecture-player-title {
+ font-size: 12px;
+ font-weight: 600;
+ opacity: 0.85;
+}
+
+.chatpapers-lecture-beat-list {
+ display: flex;
+ flex-direction: column;
+ gap: 4px;
+ max-height: 220px;
+ overflow-y: auto;
+}
+
+.chatpapers-lecture-beat-item {
+ display: flex;
+ align-items: center;
+ justify-content: space-between;
+ gap: 8px;
+ padding: 8px 10px;
+ border: 1px solid color-mix(in srgb, #1f7a8c 18%, transparent);
+ border-radius: 8px;
+ background: color-mix(in srgb, #1f7a8c 4%, transparent);
+ cursor: pointer;
+ text-align: left;
+ font: inherit;
+ color: inherit;
+}
+
+.chatpapers-lecture-beat-item:hover:not(:disabled) {
+ background: color-mix(in srgb, #1f7a8c 10%, transparent);
+}
+
+.chatpapers-lecture-beat-item.is-active {
+ border-color: color-mix(in srgb, #16a34a 45%, transparent);
+ background: color-mix(in srgb, #16a34a 10%, transparent);
+}
+
+.chatpapers-lecture-beat-item.is-pending {
+ opacity: 0.55;
+ cursor: default;
+}
+
+.chatpapers-lecture-beat-meta {
+ display: flex;
+ flex-direction: column;
+ gap: 2px;
+ min-width: 0;
+}
+
+.chatpapers-lecture-beat-type {
+ font-size: 10px;
+ font-weight: 600;
+ opacity: 0.65;
+ text-transform: uppercase;
+ letter-spacing: 0.03em;
+}
+
+.chatpapers-lecture-beat-title {
+ font-size: 12px;
+ line-height: 1.35;
+ overflow: hidden;
+ text-overflow: ellipsis;
+ white-space: nowrap;
+}
+
+.chatpapers-lecture-beat-status {
+ flex-shrink: 0;
+ font-size: 11px;
+ opacity: 0.7;
+}
+
+.chatpapers-lecture-player-controls {
+ display: flex;
+ align-items: center;
+ justify-content: center;
+ gap: 8px;
+}
+
+.chatpapers-lecture-player-btn {
+ display: inline-flex;
+ align-items: center;
+ justify-content: center;
+ width: 36px;
+ height: 36px;
+ border: 1px solid color-mix(in srgb, #1f7a8c 30%, transparent);
+ border-radius: 999px;
+ background: transparent;
+ cursor: pointer;
+ color: inherit;
+}
+
+.chatpapers-lecture-player-btn:hover {
+ background: color-mix(in srgb, #1f7a8c 12%, transparent);
+}
+
+.chatpapers-lecture-error {
+ padding: 12px;
+ color: #dc2626;
+ font-size: 12px;
+}
diff --git a/addon/content/icons/lecture.svg b/addon/content/icons/lecture.svg
new file mode 100644
index 0000000..7aa8efe
--- /dev/null
+++ b/addon/content/icons/lecture.svg
@@ -0,0 +1,28 @@
+
diff --git a/addon/content/preferences.xhtml b/addon/content/preferences.xhtml
index 03a28b1..5e3f456 100644
--- a/addon/content/preferences.xhtml
+++ b/addon/content/preferences.xhtml
@@ -116,6 +116,69 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
diff --git a/addon/locale/en-US/addon.ftl b/addon/locale/en-US/addon.ftl
index c15187c..9809c6a 100644
--- a/addon/locale/en-US/addon.ftl
+++ b/addon/locale/en-US/addon.ftl
@@ -58,3 +58,38 @@ cite-empty = No styles available
cite-copied = Bibliography copied
cite-exported = Bibliography exported
cite-export-cancel = Export cancelled
+
+item-section-lecture-head = Voice lecture
+item-section-lecture-sidenav = Voice lecture
+
+lecture-loading = Loading voice lecture…
+lecture-mode-label = Voice lecture · Prepare then listen
+lecture-empty-title = AI lecture mode
+lecture-empty-desc = One-click prep: overview and paragraph-by-paragraph listening, with typed Q&A and spoken answers.
+lecture-start = Start preparation
+lecture-status-idle = Click Start preparation — AI reads the full paper first
+lecture-status-preparing = Preparing lecture…
+lecture-status-ready = Lecture ready ✓
+lecture-status-ready-phase1 = Preparation done (Phase 1: summary ready)
+lecture-status-ready-step1 = Ready: { $count } segments — tap a segment or press Play
+lecture-status-failed = Preparation incomplete
+lecture-beat-list-title = Overview lecture
+lecture-play-start = Play
+lecture-play-pause = Pause
+lecture-play-next = Next
+lecture-play-prev = Previous
+lecture-play-complete = Playback finished
+lecture-beat-tts-pending = Synthesizing
+lecture-beat-tts-failed = Failed
+lecture-reprepare = Re-prepare
+lecture-no-pdf = No PDF attachment on this item
+lecture-attach-error = Cannot read PDF attachment
+lecture-mount-error = Failed to load voice lecture pane
+lecture-platform-hint-mac = macOS: set MinerU path under Preferences → ChatPapers → Voice lecture, or leave empty to use mineru on PATH.
+lecture-platform-hint-win = Windows: set the full path to mineru.exe under Preferences → ChatPapers → Voice lecture.
+lecture-platform-hint-other = Configure MinerU path in Preferences to enable structured PDF parsing.
+
+lecture-tts-provider = TTS: { $provider }
+lecture-tts-test = Test voice
+lecture-tts-testing = Synthesizing test speech…
+lecture-tts-test-done = Speech ready — playing
diff --git a/addon/locale/en-US/preferences.ftl b/addon/locale/en-US/preferences.ftl
index 77a5155..4586908 100644
--- a/addon/locale/en-US/preferences.ftl
+++ b/addon/locale/en-US/preferences.ftl
@@ -27,3 +27,11 @@ prefs-hint-cloud = Cloud provider: paper text is sent to the selected API endpoi
prefs-models-empty = No models returned.
prefs-models-pick = Enter a model id from the list (showing up to 30):
prefs-privacy = Privacy: ChatPapers only sends content when you click Send/Summarize, and only to the endpoint you configured. API keys are stored in local Zotero preferences.
+
+prefs-section-lecture = Voice lecture
+prefs-tts-provider = TTS engine
+prefs-tts-voice = Voice (optional)
+prefs-piper-path = Piper executable
+prefs-piper-model = Piper model path (.onnx)
+prefs-mineru-path = MinerU path
+prefs-tts-hint-empty = Choose a TTS engine. Local engines are fully offline — no API key needed.
diff --git a/addon/locale/zh-CN/addon.ftl b/addon/locale/zh-CN/addon.ftl
index 05e9c4e..c1a51bc 100644
--- a/addon/locale/zh-CN/addon.ftl
+++ b/addon/locale/zh-CN/addon.ftl
@@ -59,3 +59,37 @@ cite-copied = 参考文献已复制到剪贴板
cite-exported = 参考文献已导出
cite-export-cancel = 已取消导出
+item-section-lecture-head = 语音伴读
+item-section-lecture-sidenav = 语音伴读
+
+lecture-loading = 正在加载语音伴读…
+lecture-mode-label = 语音伴读 · 先备课再听
+lecture-empty-title = AI 老师式伴读
+lecture-empty-desc = 一键备课：整体讲解与逐段精读，随时文字提问并获得语音回答。
+lecture-start = 开始备课
+lecture-status-idle = 点击「开始备课」，AI 将先读完整篇论文
+lecture-status-preparing = 正在备课…
+lecture-status-ready = 老师已备课 ✓
+lecture-status-ready-phase1 = 备课完成（Phase 1：摘要已生成）
+lecture-status-ready-step1 = 备课完成：{ $count } 段讲解可播放，点击下方段落或播放按钮开始听课
+lecture-status-failed = 备课未完成，请重试或检查 LLM 设置
+lecture-beat-list-title = 整体讲解
+lecture-play-start = 播放
+lecture-play-pause = 暂停
+lecture-play-next = 下一段
+lecture-play-prev = 上一段
+lecture-play-complete = 本轮讲解播放完毕
+lecture-beat-tts-pending = 合成中
+lecture-beat-tts-failed = 失败
+lecture-reprepare = 重新备课
+lecture-no-pdf = 当前条目没有可用的 PDF 附件
+lecture-attach-error = 无法读取 PDF 附件
+lecture-mount-error = 语音伴读面板加载失败
+lecture-platform-hint-mac = macOS：MinerU 路径可在偏好设置 → ChatPapers → 语音伴读 中配置；留空则尝试 PATH 中的 mineru。
+lecture-platform-hint-win = Windows：请在偏好设置中填写 MinerU 可执行文件完整路径（如 mineru.exe）。
+lecture-platform-hint-other = 请在偏好设置中配置 MinerU 路径以启用 PDF 结构化解析。
+
+lecture-tts-provider = TTS 引擎：{ $provider }
+lecture-tts-test = 试听语音
+lecture-tts-testing = 正在合成测试语音…
+lecture-tts-test-done = 语音合成完成，正在播放
diff --git a/addon/locale/zh-CN/preferences.ftl b/addon/locale/zh-CN/preferences.ftl
index 4ab3ad2..8d950a3 100644
--- a/addon/locale/zh-CN/preferences.ftl
+++ b/addon/locale/zh-CN/preferences.ftl
@@ -27,3 +27,11 @@ prefs-hint-cloud = 云端 Provider：论文内容将发送到所选 API 端点
prefs-models-empty = 未返回任何模型。
prefs-models-pick = 从列表中输入模型 ID（最多显示 30 个）：
prefs-privacy = 隐私说明：仅在你点击发送/总结时，向你配置的端点发送内容。API Key 仅保存在本地 Zotero 偏好设置中。
+
+prefs-section-lecture = 语音伴读
+prefs-tts-provider = TTS 引擎
+prefs-tts-voice = 音色（可选）
+prefs-piper-path = Piper 可执行文件
+prefs-piper-model = Piper 模型路径 (.onnx)
+prefs-mineru-path = MinerU 路径
+prefs-tts-hint-empty = 请选择 TTS 引擎。本地引擎完全离线，无需 API Key。
diff --git a/addon/manifest.json b/addon/manifest.json
index 78d39d7..e0f68d8 100644
--- a/addon/manifest.json
+++ b/addon/manifest.json
@@ -14,7 +14,7 @@
"id": "__addonID__",
"update_url": "__updateURL__",
"strict_min_version": "9.0",
- "strict_max_version": "9.*"
+ "strict_max_version": "10.*"
}
}
}
diff --git a/addon/prefs.js b/addon/prefs.js
index 9cc9ccd..fd136fd 100644
--- a/addon/prefs.js
+++ b/addon/prefs.js
@@ -13,3 +13,10 @@ pref("systemPrompt", "");
pref("summaryPrompt", "");
pref("sendMetadata", true);
pref("verboseLog", false);
+pref("ttsProvider", "");
+pref("ttsApiKey", "");
+pref("ttsVoice", "");
+pref("lectureTeacherStyle", "balanced");
+pref("mineruPath", "");
+pref("piperPath", "");
+pref("piperModelPath", "");
diff --git a/docs/newidea/ARCHITECTURE.md b/docs/newidea/ARCHITECTURE.md
new file mode 100644
index 0000000..2dc2541
--- /dev/null
+++ b/docs/newidea/ARCHITECTURE.md
@@ -0,0 +1,680 @@
+# 软件架构设计文档
+
+| 字段 | 内容 |
+|------|------|
+| 产品名称 | BanDu（伴读）— Zotero 插件 |
+| 中文名 | 伴读 |
+| 英文名 | BanDu |
+| 文档版本 | v0.1 |
+| 状态 | 草案 |
+| 关联文档 | [PLAN.md](./PLAN.md)、[PRD.md](./PRD.md) |
+
+---
+
+## 1. 架构概述
+
+### 1.1 架构风格
+
+**Local-first 桌面插件架构**：核心计算与缓存均在用户机器完成；仅将段落文本与用户问题发送至用户配置的 LLM / TTS 服务。无自建后端（一期）。
+
+### 1.2 设计原则
+
+| 原则 | 说明 |
+|------|------|
+| 备课一次、缓存复用 | Lecture 持久化；attachment ID + 文件哈希为缓存键 |
+| 增量交付 | 摘要 → beats → 逐段 → TTS 分阶段就绪，缩短 Time-to-Listen |
+| UI / 领域 / 集成 分离 | 降低 Zotero API 变动影响 |
+| Grounding 优先 | 生成内容必须可追溯到段落 ID |
+| 优雅降级 | refs 无效、无 timestamps、无 BBT 均有 fallback |
+
+### 1.3 系统上下文
+
+```
+ ┌─────────────────┐
+ │ 用户 │
+ └────────┬────────┘
+ │
+ ┌──────────────▼──────────────┐
+ │ Zotero 7 (宿主) │
+ │ ┌────────────────────────┐ │
+ │ │ BanDu 插件 │ │
+ │ │ UI │ Engine │ Store │ │
+ │ └─────────┬──────────────┘ │
+ └────────────┼────────────────┘
+ │
+ ┌─────────────────┼─────────────────┐
+ │ │ │
+ ▼ ▼ ▼
+ ┌───────────┐ ┌────────────┐ ┌─────────────┐
+ │ MinerU │ │ LLM API │ │ TTS API │
+ │ (本地子进程)│ │ / Ollama │ │ ElevenLabs等 │
+ └───────────┘ └────────────┘ └─────────────┘
+ │
+ ▼
+ ┌───────────┐
+ │ SQLite + │
+ │ mp3 文件 │
+ └───────────┘
+```
+
+---
+
+## 2. 逻辑分层
+
+```
+┌─────────────────────────────────────────────────────────┐
+│ Presentation Layer（表现层） │
+│ • BanduPane（Item pane 伴读页） │
+│ • LibraryWindow（伴读库） │
+│ • SettingsPane（设置） │
+│ • PlayerController（播放状态机 +