v1.3.0

[kangfenmao]

⚠️ 仅供内部测试使用，普通用户请勿更新

⚠️ 更新务必备份数据，如果没有备份数据，升级后无法降级

What's Changed

• refactor(TopicsTab): Use onContextMenu instead of onMouseEnter by @teojs in refactor(TopicsTab): Use onContextMenu instead of onMouseEnter #5459
• Refactor/message block structure by @MyPrototypeWhat in Refactor/message block structure #4660
• feat: add painting aihubmix provider by @DDU1222 in feat: add painting aihubmix provider #4503
• feat: 添加 messageBlock、messageThunk 和 useMessageOperations 使用指南文档 by @MyPrototypeWhat in feat: 添加 messageBlock、messageThunk 和 useMessageOperations 使用指南文档 #5476
• refactor(MainTextBlock): enhance content processing by ignoring tooluse by @DeJeune in refactor(MainTextBlock): enhance content processing by ignoring tooluse #5483
• fix: message do not use knowledge by @eeee0717 in fix: message do not use knowledge #5485
• refactor: 重构网络搜索模块，支持快捷菜单快速切换 by @teojs in refactor: 重构网络搜索模块，支持快捷菜单快速切换 #5291
• perf: <tool_use> display by @koujialong in perf: <tool_use> display #5489
• fix(GeminiProvider): relocate grounding metadata handling by @DeJeune in fix(GeminiProvider): relocate grounding metadata handling #5490
• fix: The Error display of the failed mcp call shows that the Error type cannot be displayed by @koujialong in fix: The Error display of the failed mcp call shows that the Error type cannot be displayed #5492
• feat: optimize extract logic by @eeee0717 in feat: optimize extract logic #5470
• feat(messageThunk): integrate autoRenameTopic functionality to update topic names based on assistant responses by @DeJeune in feat(messageThunk): integrate autoRenameTopic functionality to update topic names based on assistant responses #5504
• fix: determining thinking process using block status by @alephpiece in fix: determining thinking process using block status #5509
• fix: update webSearch type and results structure in upgrade logic by @MyPrototypeWhat in fix: update webSearch type and results structure in upgrade logic #5512
• style: optimize mcp arg name display layout when tool or prompt descr… by @akeyz in style: optimize mcp arg name display layout when tool or prompt descr… #5467
• Fix/message block structure by @MyPrototypeWhat in Fix/message block structure #5536
• feat(WindowService): add main window monitor for renderer process events by @beyondkmp in feat(WindowService): add main window monitor for renderer process events #5532
• fix: prevent overriding block status on aborting by @alephpiece in fix: prevent overriding block status on aborting #5547
• feat: support streaming for model health check by @alephpiece in feat: support streaming for model health check #5546
• Fix error [错误]: 默认助手模型设置，当模型温度设置为0时，退出后不生效，非0值却可以。 #5456 by @nmnmtttt in Fix error #5456 #5543
• fix: refresh thinking state on streaming by @alephpiece in fix: refresh thinking state on streaming #5557
• fix(messageThunk): ensure usage is only estimated when not already present in response by @DeJeune in fix(messageThunk): ensure usage is only estimated when not already present in response #5553
• fix: reset hasReasoningContent by @alephpiece in fix: reset hasReasoningContent #5563
• feat(OpenAIProvider): support image edit by @DeJeune in feat(OpenAIProvider): support image edit #5469
• perf: improve inputbar button memorization by @alephpiece in perf: improve inputbar button memorization #5577
• fix(mcp-tools): improve error message formatting and response handlin… by @MyPrototypeWhat in fix(mcp-tools): improve error message formatting and response handlin… #5565
• refactor: enhance image block rendering and styling by @DeJeune in refactor: enhance image block rendering and styling #5567
• refactor: remove unused selection window by @0xfullex in refactor: remove unused selection window #5586
• fix: 调整图片高度，修复gemini画图 by @eeee0717 in fix: 调整图片高度，修复gemini画图 #5585
• refactor: Seperate MiniWindow loading from MainWindow by @0xfullex in refactor: Seperate MiniWindow loading from MainWindow #5581
• Feat/qwen3 support by @DeJeune in Feat/qwen3 support #5533
• feat: implement store synchronization across windows by @0xfullex in feat: implement store synchronization across windows #5592
• perf(inputbar): improve long text paste by @alephpiece in perf(inputbar): improve long text paste #5580
• refactor: simplify custom CSS functionality by store sync by @0xfullex in refactor: simplify custom CSS functionality by store sync #5596
• fix(Anthropic): add base64 file handling to IPC and file management by @DeJeune in fix(Anthropic): add base64 file handling to IPC and file management #5595
• fix: resolve scrolling issue on files page ([错误]: 文件页面的 Image 选项卡当图片过多时无法滚动 #5618) by @KumaKorin in fix: resolve scrolling issue on files page (#5618) #5619
• Remove redundant condition check in MCP service args handling by @yushihang in Remove redundant condition check in MCP service args handling #5606
• fix(i18n): add missing translation for Ru by @DeJeune in fix(i18n): add missing translation for Ru #5631
• fix(messageThunk): enhance reset message logic to include model for s… by @MyPrototypeWhat in fix(messageThunk): enhance reset message logic to include model for s… #5632
• feat(messageOperations): add editMessageBlocks functionality and upda… by @MyPrototypeWhat in feat(messageOperations): add editMessageBlocks functionality and upda… #5641
• fix(OpenAIProvider): enhance token budget calculation in OpenRouter by @DeJeune in fix(OpenAIProvider): enhance token budget calculation in OpenRouter #5625
• refactor(WindowService): improve context menu setup by @beyondkmp in refactor(WindowService): improve context menu setup  #5589
• fix the issue of webdav local backup file storage by @sun11 in fix the issue of webdav local backup file storage #5643
• refactor(WindowService, ThemeProvider): streamline title bar style and theme toggling logic by @beyondkmp in refactor(WindowService, ThemeProvider): streamline title bar style and theme toggling logic #5633
• fix(models): normalize model IDs to lowercase for consistency by @DeJeune in fix(models): normalize model IDs to lowercase for consistency #5642
• chore(dependencies): update electron-updater to version 6.6.4 and remove patches by @beyondkmp in chore(dependencies): update electron-updater to version 6.6.4 and remove patches #5650
• chore(dependencies): update electron-builder to 26.0.15 by @beyondkmp in chore(dependencies): update electron-builder to 26.0.15 #5651
• fix: remove redundant local variables by @Pleasurecruise in fix: remove redundant local variables #5654
• feat: 支持自定义助手地址 by @CreatorZZY in feat: 支持自定义助手地址 #5540
• Fix/image height by @DeJeune in Fix/image height #5658
• fix: conditionally show loading more spinner by @alephpiece in fix: conditionally show loading more spinner #5670
• feat: popup question editor support translation assistant by @chenxi-null in feat: popup question editor support translation assistant #5660
• fix: infinite scroller layout by @alephpiece in fix: infinite scroller layout #5671
• refactor: add Linux support for margin adjustments in MinappPopupConainer and McpSettingsNavbar by @beyondkmp in refactor: add Linux support for margin adjustments in MinappPopupConainer and McpSettingsNavbar #5673
• Fix: 'Web Search' and 'Clear Context' don't work by @chenxi-null in Fix: 'Web Search' and 'Clear Context' don't work #5677
• feat(OpenAI): new Responses API support by @DeJeune in feat(OpenAI): new Responses API support #5621
• fix: openaiprovider timeout is not an integer by @Pleasurecruise in fix: openaiprovider timeout is not an integer #5681
• fix: include thinking_millsec in message block creation by @DeJeune in fix: include thinking_millsec in message block creation #5685
• fix: memorize assistant with model by @alephpiece in fix: memorize assistant with model #5701
• feat(miniapps): add Grok / X by @ipcjs in feat(miniapps): add Grok / X #5706
• feat: emit event on message complete by @0xfullex in feat: emit event on message complete #5696
• fix: update thinkingbutton model checks by @DeJeune in fix: update thinkingbutton model checks  #5708
• feat(WebSearchButton): add 'Disable Web Search' option by @teojs in feat(WebSearchButton): add 'Disable Web Search' option #5717
• fix(StoreSyncService): set flag bug when register IPC handler by @0xfullex in fix(StoreSyncService): set flag bug when register IPC handler #5715
• 优化消息块渲染性能，使用 useMemo 缓存图像块组；移除调试日志以清理代码 by @MyPrototypeWhat in 优化消息块渲染性能，使用 useMemo 缓存图像块组；移除调试日志以清理代码 #5722
• feat: support dashscope reranker by @eeee0717 in feat: support dashscope reranker #5725
• fix: chat message translate by @Pleasurecruise in fix: chat message translate #5682
• feat: change citation list style by @Pleasurecruise in feat: change citation list style #5516
• feat: optimize knowledge recognize by @eeee0717 in feat: optimize knowledge recognize #5707
• fix: knowledge base url error by @Pleasurecruise in fix: knowledge base url error #5735
• refactor(middleware): add extractReasoningMiddleware for enhanced text processing by @DeJeune in refactor(middleware): add extractReasoningMiddleware for enhanced text processing #5637
• fix: failed to connect OAuth MCP server by @chenxi-null in fix: failed to connect OAuth MCP server #5709
• fix: citation list loading by @Pleasurecruise in fix: citation list loading #5742
• feat: Add OpenDAL based Remote Storage class by @Xuanwo in feat: Add OpenDAL based Remote Storage class #2700
• feat: add zoom factor setting and localization support by @beyondkmp in feat: add zoom factor setting and localization support #5665
• chore: remove unuseful doc by @DeJeune in chore: remove unuseful doc #5740
• feat[MCP]: implement login shell environment retrieval by @vaayne in feat[MCP]: implement login shell environment retrieval #5739
• feat(MCP): update auto-install server name and add server registratio… by @MyPrototypeWhat in feat(MCP): update auto-install server name and add server registratio… #5751
• fix(theme): citation list dark theme display & url repeatedly requesting by @Pleasurecruise in fix(theme): citation list dark theme display & url repeatedly requesting #5752
• feat: enhance search functionality with optional HTTP options by @MyPrototypeWhat in feat: enhance search functionality with optional HTTP options #5765
• Feat: change Gemini Safety Settings to OFF; upgrade @google/genai package by @ous50 in Feat: change Gemini Safety Settings to OFF; upgrade @google/genai package #5763
• fix(MCP): Revise the server type judgment logic to ensure the correct… by @shlroland in fix(MCP): Revise the server type judgment logic to ensure the correct… #5769
• Feat/claude websearch support by @DeJeune in Feat/claude websearch support #5771
• feat: Custom mini app by @CreatorZZY in feat: Custom mini app #5731
• refactor(ModelList): improve model list style and grouping by @alephpiece in refactor(ModelList): improve model list style and grouping #5674
• feat(minapps): add n8n mini app with logo and URL. by @CreatorZZY in feat(minapps): add n8n mini app with logo and URL. #5776
• fix: display image from mcp response by @kanweiwei in fix: display image from mcp response #5780
• feat: tooluse result display style optimization by @koujialong in feat: tooluse result display style optimization #5758
• fix: remove useless border and border radius for code block by @kovsu in fix: remove useless border and border radius for code block #5773
• perf: virtual list for quick panel and SelectModelPopup by @alephpiece in perf: virtual list for quick panel and SelectModelPopup #5594
• feat: add the "Export Agent" feature by @shlroland in feat: add the "Export Agent" feature #5789
• fix: user message usage by @alephpiece in fix: user message usage #5657
• feat: ParateraAI 添加支持 by @CreatorZZY in feat: ParateraAI 添加支持 #5792
• fix: SelectModelPopup sticky header by @alephpiece in fix: SelectModelPopup sticky header #5795
• feat: support both function call and system prompt for MCP tools by @kanweiwei in feat: support both function call and system prompt for MCP tools #5499
• fix: update input schema reference in MessageTools for accurate tool … by @kanweiwei in fix: update input schema reference in MessageTools for accurate tool … #5804
• feat: add inspect option to context menu with localization support by @beyondkmp in feat: add inspect option to context menu with localization support #5807
• feat(Qwen3): Add Qwen3 Model Thinking Mode Switch by @jwcrystal in feat(Qwen3): Add Qwen3 Model Thinking Mode Switch #5781
• feat: combine to general reranker by @eeee0717 in feat: combine to general reranker #5818
• fix: SelectModelPopup scrolling behaviour by @alephpiece in fix: SelectModelPopup scrolling behaviour #5812
• fix: qwen3 check by @DeJeune in fix: qwen3 check #5811
• feat(webSearch): add Bocha web search provider integration by @DeJeune in feat(webSearch): add Bocha web search provider integration #5608
• docs(README): add Architecture overview document by @jwcrystal in docs(README): add Architecture overview document #5824
• fix: update selectedTypes logic in ModelEditContent for better handling by @kanweiwei in fix: update selectedTypes logic in ModelEditContent for better handling #5819
• feat: add prompt display control by @985563349 in feat: add prompt display control #5439
• style: remove small size prop from Switch component in AssistantModel… by @kanweiwei in style: remove small size prop from Switch component in AssistantModel… #5833
• feat: Agent Regular Phrases by @CreatorZZY in feat: Agent Regular Phrases #5775
• feat: regularPhrases可以随Agent一起导出。 by @CreatorZZY in feat: regularPhrases可以随Agent一起导出。 #5836
• fix: jina embedding error by @DeJeune in fix: jina embedding error #5839
• feat(i18n): 更新机器翻译多语言，使用Qwen3 236b进行机器翻译。 by @CreatorZZY in feat(i18n): 更新机器翻译多语言，使用Qwen3 236b进行机器翻译。 #5840
• fix: check isComposing on keydown by @alephpiece in fix: check isComposing on keydown #5848
• fix: clean up code by @Pleasurecruise in fix: clean up code #5851
• revert: customtag tooltip delay by @alephpiece in revert: customtag tooltip delay #5856
• Revert "fix: update selectedTypes logic in ModelEditContent for better handling" by @DeJeune in Revert "fix: update selectedTypes logic in ModelEditContent for better handling" #5858
• fix: conditionally aquire focus for inputbar by @alephpiece in fix: conditionally aquire focus for inputbar #5860
• refactor(checkAPI): check api or model with stream enabled first by @alephpiece in refactor(checkAPI): check api or model with stream enabled first #5857
• feat: add message translate copy & close by @Pleasurecruise in feat: add message translate copy & close #5684
• fix(Inputbar): do not reset selection on focus by @alephpiece in fix(Inputbar): do not reset selection on focus #5866
• feat: add motion library for animations and enhance Spinner and Messa… by @MyPrototypeWhat in feat: add motion library for animations and enhance Spinner and Messa… #5869
• fix: next release bugs by @MyPrototypeWhat in fix: next release bugs #5801

New Contributors

• @nmnmtttt made their first contribution in Fix error #5456 #5543
• @KumaKorin made their first contribution in fix: resolve scrolling issue on files page (#5618) #5619
• @yushihang made their first contribution in Remove redundant condition check in MCP service args handling #5606
• @sun11 made their first contribution in fix the issue of webdav local backup file storage #5643
• @CreatorZZY made their first contribution in feat: 支持自定义助手地址 #5540
• @ipcjs made their first contribution in feat(miniapps): add Grok / X #5706
• @Xuanwo made their first contribution in feat: Add OpenDAL based Remote Storage class #2700
• @ous50 made their first contribution in Feat: change Gemini Safety Settings to OFF; upgrade @google/genai package #5763
• @kovsu made their first contribution in fix: remove useless border and border radius for code block #5773
• @jwcrystal made their first contribution in feat(Qwen3): Add Qwen3 Model Thinking Mode Switch #5781
• @985563349 made their first contribution in feat: add prompt display control #5439

Full Changelog: v1.2.10...v1.3.0

---

This discussion was created from the release v1.3.0.

[hang333]

o3 模型 with fc mcp 出现错误

[DeJeune]

这个是没有做schema校验，临时解决方案是退回到prompt

[hang333]

这个是没有做schema校验，临时解决方案是退回到prompt

尴尬的是，o3 模型根本不吃 prompt，所以我一直在等 fc 回归hh

[DeJeune]

如果mcp是你自己写的，可以改一下函数一些字段

[CreatorZZY]

OS: Ubuntu 24.04
文件：Cherry-Studio-1.3.0-x86_64.AppImage

[luosc]

试试这个：https://askubuntu.com/questions/1524005/appimages-wont-run-on-24-04

[CreatorZZY]

以前不需要这个功能的啊

[gonisan]

启动速度对比1.2.10明显变慢
os： macOS Sequoia 15.4.1 arm64

[WJG-fefi]

知识库重排序仍然不能用；菜单显示不全：知识库、思维链、网络

[eeee0717]

用的什么重排模型？
报错信息是什么？

[WJG-fefi]

用的什么重排模型？ 报错信息是什么？

重排用的硅基流动：BAAI/bge-reranker-v2-m3。没有报错，只是大模型回答相同的问题，使用重排答非所问，不使用则可以正常问答。1.3.1测试，相同的问题，使用重排的情况下，回答有时候是知识库内容，有时候不是，相比其他版本有所改进。但这之前的版本在使用重排模型就完全是非答所问。不清楚是重排模型的问题，还是版本的问题。

[WJG-fefi]

设置→常规设置→代理模式，设置为“不使用代理”则可正常使用重排序模型

[FLC-ytfl]

输入框上拉后，旁边的这个上下按键就按不到了

[FLC-ytfl]

好像一直都有这个问题，之前提过issue也没解决，虽然影响不大

[ous50]

留意到如果模型没有返回或者中途停止了的话，think time 貌似不会停下来。
可能会导致内存泄漏（？）

[kawainotu]

#5914 似乎在修了

[DeJeune]

#5914 似乎在修了

测试一下，还会出现吗

[hutchisr]

Using Claude 3.7 Sonnet with both "Stream output" and "Enable tool use" enabled causes error Cannot read properties of undefined (reading '0') when calling functions. Disabling either option (but keeping the other enabled) works as expected.