v0.12.2

• fix: Fixed the issue of audio duration loss in recorded files
• fix: Fixed the issue with low recording volume on Safari browser
• fix: Fixed the issue where the access password box in docker mode was hidden by mistake

---

• fix: 修复录制文件中音频时长丢失的问题
• fix: 修复 Safari 浏览器下录制音量低的问题
• fix: 修复 docker 模式下访问密码框被误隐藏的问题

v0.12.0

The new version uses Tauri to develop a cross-platform application client that supports a permanent menu bar, doubling your work efficiency~

The current version has truly realized conversations with Gemini. The new version uses the audio recognition capabilities of the Gemini 1.5 series, without the need for the browser's voice recognition function, to directly send the original audio to Gemini, allowing Gemini to "understand" your words. This mode is similar to the Gemini Live feature demonstrated at Google I/O 2024.

---

新版本使用 Tauri 开发了跨平台的应用客户端，支持常驻菜单栏，让您的工作效率翻倍～

当前版本已经真正意义上实现了与 Gemini 对话。新版本利用 Gemini 1.5 系列的音频识别能力，无需借助浏览器的语音识别功能，直接将原始音频发送给 Gemini，让 Gemini 可以"听懂"你的话。这种模式类似于 Google I/O 2024 上演示的 Gemini Live 功能。

v0.11.8

• feat: Completed the development of desktop applications based on tauri
• feat: Use Gemini 1.5 audio recognition capabilities to replace traditional TTS
• feat: Added Recorder tool class for user recording
• feat: Rewrite the strong rule to adapt to Gemini generation grammar
• refactor: Refactored the message rendering module and optimized the image and audio rendering logic
• refactor: Hide the Copy and Speak buttons when there is no text
• refactor: Hide access password field when no password protection is enabled
• build: Add Github Actions and issue template
• build: Add docker-compose.yml
• build: Automatic cross-platform compilation using Tauri Action
• doc: Added docker-compose.yml deployment method to the documentation and removed some questions from FAQ
• fix: Add a background color to the bottom function bar to prevent the text from overlapping with the bottom function bar
• fix: Fixed the problem that responsive pop-up windows do not work on mobile pages

---

• feat: 完成基于 tauri 的桌面应用开发
• feat: 使用 Gemini 1.5 的音频识别能力替代传统 TTS
• feat: 增加 Recorder 工具类，用于用户录音
• feat: 重写 strong 规则，适配 Gemini 生成语法
• refactor: 重构消息渲染模块，优化图片和音频渲染逻辑
• refactor: 无文字时隐藏复制和朗读按钮
• refactor: 无密码保护时隐藏访问密码字段
• build: 增加 Github Actions 和 issue 模板
• build: 增加 docker-compose.yml
• build: 使用 Tauri Action 自动跨平台编译
• doc: 文档增加 docker-compose.yml 部署方式，FAQ 移除部分问题
• fix: 底部功能栏增加背景色，避免文字与底部功能栏重叠
• fix: 修复响应式弹窗在移动端页面无法使用的问题

v0.10.8

• feat: Add PWA installation button
• feat: Added file upload size limit
• refactor: Optimize PWA file generation logic
• fix: Disable modification of proxy api url without apikey
• doc: Modify app screenshot and add NEXT_PUBLIC_UPLOAD_LIMIT environment variable description
• doc: Added issues and solutions regarding the inability of vercel's default domain name to be accessed normally in China

---

• feat: 添加 PWA 安装按钮
• feat: 添加文件上传大小限制
• refactor: 优化 PWA 文件生成逻辑
• fix: 禁止在没有 apikey 的情况下修改代理 api 链接
• doc: 修改应用截图并添加NEXT_PUBLIC_UPLOAD_LIMIT环境变量描述
• doc: 新增 vercel 默认域名在国内无法正常访问的问题及解决办法

v0.10.7

• feat: Added text tools, including regenerate, edit, delete, copy and speak
• feat: Supports pasting and uploading images/files from the clipboard
• feat: Supports drag and drop upload of images/files
• feat: Added the function of uploading large files in parts
• feat: Support PWA (experimental)
• feat: Add custom model list and custom default model
• feat: Added model parameter setting page
• feat: Assistant market adds classification filtering function
• feat: Forward data using the common proxy interface
• feat: Added error message display and file cleaning for file upload
• feat: Added file upload proxy url configuration parameters
• refactor: Reconstruct file upload logic to be compatible with docker deployment solution
• fix: Fixed the problem of confusing page layout when the code field is too long
• chore: Added file upload proxy configuration
• build: Use multiple rewrites rule hack to solve the problem of invalid rewrites rule variables in standalone mode

---

• feat: 添加文本操作工具，包括重新生成、编辑、删除、复制和朗读
• feat: 支持从剪贴板粘贴和上传图像/文件
• feat: 支持拖放上传图像/文件
• feat: 新增大文件分段上传功能
• feat: 支持 PWA（实验性）
• feat: 添加自定义模型列表和自定义默认模型
• feat: 新增模型参数设置页面
• feat: 助手市场增加分类筛选功能
• feat: 使用通用代理接口转发数据
• feat: 添加了文件上传的错误消息显示和文件清理
• feat: 添加文件上传代理url配置参数
• refactor: 重构文件上传逻辑，兼容docker部署方案
• fix: 修复代码字段过长时页面布局混乱的问题
• chore: 添加文件上传代理配置
• build: 使用多个重写规则 hack 解决独立模式下重写规则变量无效的问题

v0.9.6

• feat: Add system instruction support
• feat: Support media files as prompt content
• feat: Chat UI supports attachments, Complete file upload function
• feat: Added server-side reverse proxy (experimental)
• feat: Support model settings
• feat: Since the data storage structure has changed since version 0.9.3, a new data migration function has been added
• refactor: Reconstruct the topic square and introduce assistant market
• refactor: Refactoring Assistant Recommendations
• refactor: Reconstruct the settings panel to optimize page layout on small screens
• refactor: The message list uses the official data format
• refactor: Optimize the display logic of error message
• refactor: Optimize the processing logic of visual model and text model
• refactor: Use localforage to replace the localStorage implementation
• refactor: Hide module in some browsers where talk mode is not available
• refactor: Optimize access password processing logic and remove NEXT_PUBLIC_ENABLE_PROTECT variable
• doc: Update the README document, add some common problems and solutions, and add more function screenshots
• doc: Added Cloudflare Worker creation guide
• fix: Fix speech synthesis initialization parameter settings
• fix: Fixed an issue where regenerated answers could not take effect in some scenarios
• chore: Some page components have been changed to lazy loading to improve the loading speed of the home page.

---

• feat：添加系统指令支持
• feat：支持媒体文件作为提示内容
• feat：聊天 UI 支持附件，完成文件上传功能
• feat：添加了服务器端反向代理（实验性）
• feat：支持模型设置
• feat: 由于 0.9.3 版本以来数据存储结构发生变化，增加数据迁移功能
• refactor：重构主题广场，引入助理市场
• refactor：重构助理推荐功能
• refactor：重构设置面板，优化小屏幕上的页面布局
• refactor：消息列表使用官方数据格式
• refactor：优化错误信息的显示逻辑
• refactor：优化视觉模型和文本模型的处理逻辑
• refactor：使用 localforage 替换 localStorage 实现
• refactor：在某些无法使用对话模式的浏览器中隐藏功能模块
• refactor：优化访问密码处理逻辑，删除 NEXT_PUBLIC_ENABLE_PROTECT 变量
• doc：更新 README 文档，添加一些常见问题及解决方案，并添加更多功能截图
• doc：添加 Cloudflare Worker 创建指南
• fix：修复语音合成初始化参数设置错误的问题
• fix：修复部分场景下重新生成答案无法生效的问题
• chore：部分页面组件改为延迟加载，提高首页加载速度。

v0.8.2

• chore: Use the official npm package @google/generative-ai to replace third-party npm packages to improve security
• fix: Fixed the issue where the theme switching interface is invalid
• fix: Fix request permission verification under password protection
• fix: Fixed the issue where the default synthesized sound source was missing
• fix: Fixed the issue where a request would still be generated if the access password does not exist
• fix: Fix the abnormal compatibility issue between NEXT_PUBLIC_ENABLE_PROTECT and ACCESS_PASSWORD parameters in the server api

---

• chore：使用官方npm包@google/generative-ai替换第三方npm包以提高安全性
• fix：修复主题切换界面无效的问题
• fix：修复密码保护下的请求权限验证
• fix：修复默认合成音源丢失的问题
• fix：修复访问密码不存在仍会生成请求的问题
• fix：修复服务器api中NEXT_PUBLIC_ENABLE_PROTECT和ACCESS_PASSWORD参数兼容性异常的问题

v0.8.0

Only for static deployment

仅用于静态部署