版本 0.28：导出稳定性与内容完整性全面升级

本次更新聚焦“复杂页面可导出、长会话更稳定、Word 显示更自然”。

ChatGPT Deep Research 导出能力增强
新增深度研究页面 iframe 桥接抓取机制，支持跨框架内容提取。
优化延迟加载与虚拟化场景下的抓取逻辑，减少“只导出占位内容/正文缺失”问题。
增加来源校验、超时与重试策略，提升安全性与成功率。
Word / PDF / HTML 导出质量提升
修复 Word 分隔线异常问题（长方形块、双线、偶发不可见），统一为更稳定的浅灰单线样式。
优化代码块、表格、数学公式（KaTeX）与图文混排的导出还原效果。
调整打印分页策略，减少长对话导出时的空白页、内容断裂和错位。
ChatGPT / Claude 解析链路优化
改进复杂消息结构解析，提升长文本与结构化内容提取完整度。
强化图片与富文本后处理，降低导出后的样式丢失与布局异常概率。
优化推理/思考内容处理路径，确保按用户设置准确保留或排除。
稳定性与容错增强
导出流程补充更多异常兜底分支，覆盖超时、上下文失效、回调失败等边界情况。
提升复制导出与离屏导出链路鲁棒性，降低偶发失败率。

Version 0.28: Major upgrades to export stability and content completeness

This release focuses on reliable exports for complex pages, better long-conversation handling, and cleaner Word output.

Enhanced ChatGPT Deep Research export
Added an iframe bridge extraction path for deep research pages.
Improved handling of delayed rendering and virtualized content to reduce missing-body exports.
Added source validation, timeout handling, and retry logic for better safety and reliability.
Improved Word / PDF / HTML export quality
Fixed Word divider rendering issues (rectangle-like blocks, double-line look, occasional invisible lines), now using a more stable subtle single-line style.
Improved fidelity for code blocks, tables, KaTeX math, and mixed media layouts.
Refined pagination to reduce blank pages, content breaks, and layout shifts in long exports.
Better parsing pipeline for ChatGPT / Claude
Improved parsing of complex message structures for more complete content capture.
Strengthened image and rich-text post-processing to reduce style/layout regressions.
Improved reasoning/thinking content handling to honor user export settings more accurately.
Stronger resilience and fallback behavior
Added broader fallback branches for timeout, invalid context, and callback failure scenarios.
Increased robustness of copy-export and offscreen export pipelines.

v0.0.28 紧急更新说明（应用商店文案）

中文

这次是一次紧急稳定性更新，主要用于提升导出成功率和整体使用体验。

本次更新内容

• 修复了少数情况下导出失败或中断的问题。
• 优化了大体量聊天记录导出时的稳定性和流畅度。
• 优化了下载与打印流程，减少“点了没反应”或提示不清晰的情况。
• 提升了在不同浏览器环境下的一致性与兼容性。
• 修复若干已知问题，整体导出体验更可靠。

升级建议

建议所有用户尽快升级到 0.0.28，以获得更稳定的导出体验。



English

This is an emergency stability update focused on improving export reliability and overall user experience.

What’s improved

• Fixed edge cases where exports could fail or stop unexpectedly.
• Improved stability and smoothness when exporting large conversations.
• Improved download and print behavior with clearer feedback.
• Better consistency and compatibility across different browser environments.
• Fixed several known issues to make exports more reliable overall.

Recommended action

We strongly recommend all users upgrade to 0.0.28 for a more stable export experience.

v0.0.27 新增功能（中文）
🇨🇳 中文版 (Chinese Version)
🚀 重大架构升级：极致性能与抗崩溃引擎重构
本次更新深入底层逻辑，彻底重写了数据抓取与处理管线。专为应对动辄十万字以上的“深度思考”大模型对话而生。

🛠️ 核心性能提升 (Performance & Stability)

消灭浏览器卡死 (Time Slicing)： 重构了底层数据收割机 (Harvester)。引入了全新的“时间切片”与异步微任务调度机制。现在，即使一次性导出数百轮、包含海量代码块的长对话，主页面依然如丝般顺滑，彻底告别浏览器假死与“页面无响应”弹窗。

内存暴降 80% (Memory Optimization)： 针对超长对话导出极易引发的内存溢出（OOM 崩溃）问题，采用了极端的 DOM 节点复用与显式垃圾回收 (GC) 策略。大幅降低了 V8 引擎的峰值内存占用。

Offscreen 独立沙盒引擎： 将耗费 CPU 的繁重任务（如 Word 文档打包压缩、Markdown 解析）全面迁移至独立的后台 Offscreen 环境中执行，真正实现了计算与阅读体验的物理隔离。

重构 IPC 数据流转： 优化了跨进程通信机制，摒弃了旧版全量 Base64 字符串的低效传输，改用分片与轻量级引用流转，极限提升大型文件的导出速度。

🧹 体验优化与清理 (Cleanup & Polish)

大扫除行动： 大刀阔斧地移除了旧版本中的冗余轮询器 (Polling)、过时的正则表达式，以及数十条废弃的平台样式选择器，插件包体更加纯粹轻量。

更精准的 DOM 捕获： 增强了语义化匹配规则，面对主流 AI 平台频繁的 UI 改版，提取逻辑更具弹性和容错率。
v0.0.27 What’s New (English)
🇬🇧 英文版 (English Version)
🚀 Major Architecture Update: The Extreme Performance Engine
This update brings a massive under-the-hood refactoring to our data pipeline, specifically engineered to handle massive, 100,000+ word "deep thinking" AI conversations without breaking a sweat.

🛠️ Performance & Stability

Zero UI Freezing (Time Slicing): Completely rebuilt the core Harvester. By introducing advanced "Time Slicing" and asynchronous micro-task scheduling, exporting hundreds of conversation turns is now buttery smooth. Say goodbye to browser freezes and "Page Unresponsive" warnings.

80% Peak Memory Reduction: Fixed Out-Of-Memory (OOM) crashes during massive exports. We implemented an aggressive DOM container recycling and explicit Garbage Collection (GC) strategy, drastically reducing the memory footprint during heavy document processing.

Offscreen Sandbox Processing: Heavy CPU tasks, such as generating and zipping Word (.docx) documents, have been entirely migrated to a secure, isolated Offscreen background environment, ensuring zero impact on your active reading experience.

Optimized IPC Data Flow: Replaced the inefficient, massive Base64 string transfers with a streamlined, chunked data pipeline, significantly boosting the export speed for large files containing images and code.

🧹 Cleanup & Polish

Tech Debt Purge: Aggressively removed legacy polling intervals, outdated regex parsers, and redundant CSS selectors. The extension is now noticeably lighter and cleaner.

Resilient DOM Capture: Enhanced the semantic matching rules. The export engine is now highly resilient to frequent UI changes across major AI platforms, ensuring consistent and accurate data extraction.

v0.0.26 新增功能（中文）

新增 AI 内嵌浏览模式（AI iframe）：可在侧边栏内直接打开常用 AI 页面，减少来回切换标签页。
新增 Side Panel 一体化体验：聊天查看、历史记录、设置调整都可在同一侧边栏完成。
新增 跨平台快速切换：在多个 AI 平台之间切换更顺畅。
新增 导出历史统一管理：历史记录查看更稳定、回看更方便。
新增 样式问题反馈入口：可直接反馈页面布局/显示问题。
新增 32 种语言完善支持：多语言使用体验进一步统一。
v0.0.26 What’s New (English)

Added AI embedded browsing (AI iframe): open supported AI pages directly in the side panel, with fewer tab switches.
Added a unified Side Panel workflow: chats, history, and settings are all managed in one place.
Added faster cross-platform switching across supported AI services.
Added unified export history management for more reliable viewing and easier review.
Added a layout issue feedback entry for quick visual bug reporting.
Added improved 32-language coverage for a more consistent global experience.

[English] 🚀 Major Platform Expansion! We have significantly expanded our support list. You can now export chats from 5 new major AI platforms:

Grok (xAI)

Perplexity (AI Search)

Qwen / Tongyi Qianwen (Aliyun)

Yuanbao (Tencent)

Doubao (ByteDance)

✨ Improvements:

Popup UI: Optimized the extension popup for a smoother experience.

Stability: Added comprehensive snapshot tests to ensure accurate parsing across all supported platforms.

[中文] 🚀 重磅更新：新增 5 大 AI 平台支持！ 我们极大地扩展了支持列表，现在支持从以下平台导出对话：

Grok (xAI)

Perplexity (AI 搜索)

通义千问 Qwen (阿里云)

腾讯元宝 Yuanbao

豆包 Doubao (字节跳动)

✨ 优化与改进：

界面升级：优化了插件弹窗 (Popup) 的交互体验。

稳定性提升：底层增加了大量自动化测试与 DOM 快照对比，确保在不同平台下的导出准确性。

【v0.24 Major Update / 重大更新】

🎉 New Features / 新功能:
1. Support "Deep Research" Export: Perfectly export OpenAI's Deep Research reports, preserving all tables and citation links.
(完美支持 ChatGPT Deep Research 深度研究报告导出，保留表格与引用链接。)
2. PDF Export (High Quality): New "Print to PDF" mode support. Get vector-quality PDF reports with clean layout.
(新增高清 PDF 导出模式，通过打印预览生成矢量级文档，排版更纯净。)
3. Document View Mode: Switch between "Chat Bubble" and "Document" views. Ideal for reading long reports.
(新增“文档视图”模式：一键切换气泡/文档模式，专为长文阅读设计。)
4. Image Offline Support: Images are now converted to Base64. No more broken images in your exported HTML!
(图片离线化支持：自动将图片转为 Base64，彻底解决导出后图片失效的问题。)
5. Fold/Expand Messages: Easily collapse long AI responses to keep your view clean.
(新增消息折叠功能：长对话一键折叠，让上下文浏览更清爽。)

🚀 Improvements / 优化:
- Performance Boost: Fixed browser freezing issues when exporting super long conversations (100+ turns).
(性能飞跃：修复了超长对话导致的浏览器卡顿问题，导出速度提升 10x。)
- Multi-language Support: Added support for French, Thai, Japanese, and more.
(多语言支持：新增法语、泰语、日语等本地化适配。)

🐛 Fixes / 修复:
- Fixed CSS layout distortions in certain markdown tables.
- Fixed an issue where the export button failed to mount on some pages.

Ver 0.0.23 Update Highlights:

[New] Added "Back to Top" and "Scroll to Bottom" buttons.

[New] Added image export capability for Gemini & ChatGPT.

[Improved] Optimized theme styling for exported HTML files.

[Improved] Refined general UI style details.

[Fixed] Corrected AI platform logo display errors.

[Fixed] Resolved the incorrect mounting position of the export button.

v0.0.21 Changelog
[Feature] Added support for ChatGPT Deep Research model export.
[Fix] Resolved UI freezing/hanging issues when exporting long conversations.
[Fix] Improved Keyword Extraction algorithm (reduced noise from code blocks).
[Fix] Fixed CSS layout distortions in generated HTML.
[Fix] Solved the issue where the Export Button failed to mount on DOM.

Fixed: Resolved an issue where images in Gemini exports were broken or missing.
Improved: Enhanced export stability for long conversations.