--- New Features --- General - Support resizing the iPhone app → Full support for iPhone Duo is planned for a future update. - Add action buttons to the toolbar on the Server and Model screens → You can now access context menus without having to right-click or long-press list items. Chat - Support attaching text files → You can now attach various text files, including .txt and .md, to all models that support chat. - Support attaching PDF files → You can now attach PDF files to all models that support chat. In addition, for models that support vision, each page is also sent as an image, allowing them to recognize images, diagrams, and layouts in addition to text. - Support selecting thinking levels → You can now select thinking levels from "Off", "Low", "Medium", "High", and "Max" for thinking-capable models. In addition, "None" has been renamed to "Default". --- Bug Fixes and Improvements --- General - Refresh the app icon → Redesigned the app icon and the design displayed on the iOS launch screen. - Fix issue where the window position and size could be reset on macOS - Improve the appearance of the license screens - Improve the host input on the server add/edit screen and the model name input on the add model screen so spelling is not autocorrected - Change default API timeout to 5 minutes → Reduced the likelihood of timeout errors when loading large models or when prompt processing takes longer. Model - Fix issue where the modified date of models was not displayed correctly depending on the environment - Improve verification display after model download completion Chat and Image Generation - Fix issue where the model picker could sometimes be empty on macOS Golden Gate - Fix issue where picker contents could shrink on macOS Tahoe or later - Improve the appearance of the message input area on macOS Golden Gate - Improve the appearance of the scroll to bottom button - Improve handling so that entered text and attached files are not reset when clearing chat or image generation history - Improve auto-scroll behavior Chat - Fix issue where thinking messages were not displayed when "Thinking" was set to "None" - Fix issue where scrolling would sometimes jump when streaming completed - Fix issue where files could not be dropped onto attached file icons - Fix issue where attaching images could cause performance to degrade - Fix issue where empty message bubbles could remain after stopping a chat, then editing and resending a message Image Generation - Fix issue where images would shrink when the scroll to bottom button appeared Release Notes: https://mocolamma.taikun.design/release-notes

--- New Features --- Model - Support audio model tag - Add a feature that allows scrolling by dragging tags in the model inspector - Add the ability to load and unload models from the context menu of each model → You can load models by specifying a Keep Alive time, or unload them. You can also check the loading status in the model list. - Add a feature to cancel model downloading Chat and Image Generation - Add options to set the order of models → By default, the order displayed on the model screen is used, but you can change it to your preferred order in the settings. - Add a button to scroll to the bottom → A button now appears that allows you to scroll straight to the bottom when you are outside the auto-follow scroll range. --- Bug Fixes and Improvements --- General - Fix issue where sheet backgrounds appear slightly green in iPadOS 26 Dark Mode - Fix issue where performance significantly degrades when voice control is enabled - Improve the layout of the settings screen Server - Improve the display of connection status in server inspector Model - Fix issue where the model filter icon was not displayed in macOS Sequoia - Fix issue where models could not be added in a specific format → Fixed an issue where models could not be downloaded if the model name contained a "/" in the ollama command format. - Improve model loading process → Performance has improved by automatically loading models when the app launches and when the server changes, thereby eliminating unnecessary model loading. - Improve the feature to allow copying the model name using a keyboard shortcut → You can copy the model name by pressing ⌘ (Command) + C while the model is selected. On macOS, you can also copy it by selecting “Copy” from the “Edit” menu. Chat and Image Generation - Fix issue where the corner radius of the message input field was shifted - Fix issue where auto-follow scrolling would sometimes jumps - Fix issue where the keyboard sometimes did not close when the background was tapped - Improve the model picker display while loading models - Improve the appearance of the model picker → On macOS, the width now remains fixed, and on visionOS, model names are now displayed just like on macOS. - Improve the appearance of the message input field → When an input field is disabled, it is now semi-transparent to make its disabled state easier to recognize. - Improve to scroll when editing or retrying a message Chat - Fix issue where list bullets were not displayed - Fix issue where text selection did not work properly in visionOS - Improve the display of the thinking text to use Markdown - Improve Markdown styling - Improve the performance of the thinking process Release Notes: https://mocolamma.taikun.design/release-notes

New Features - Add a copy button to code blocks in the chat screen → You can copy only the code within the code block. Bug Fixes and Improvements - Fix issues where text selection could not cross code blocks and where multiple word selection was not possible by holding down the double-click button in the chat screen - Fix issue where text selection could not be released in the chat screen on platforms other than macOS Highlights of New Features in Version 1.2.0 - Revamping the Markdown rendering engine → This incorporating chat improvements such as enhanced performance and syntax highlighting for code blocks, resulting in a more user-friendly experience. → As a result, Mocolamma now requires iOS / iPadOS 18.0 or later, macOS 15.0 or later, visionOS 2.0 or later. Please note that users on older operating systems will need to perform a software update. - Native support for visionOS - Support for attaching images in chat - Support for Ollama's image generation feature (beta) → As Ollama's image generation feature is currently in beta, future Ollama updates may introduce changes to its functionality (Ollama API), potentially causing the image generation feature to stop working in Mocolamma. - Add Image Preview feature → You can preview attached images or generated images in a large enlarged view. - Add an icon customization feature to the add/edit server screen - Add an auto-follow scroll feature to the chat screen and image generation screen - Add many advanced options to custom settings for chat - Add Arabic (ar), Chinese (Hong Kong) (zh-HK), Chinese (Simplified) (zh-Hans), Chinese (Traditional) (zh-Hant), Korean (ko), Russian (ru), Ukrainian (uk) localization by generative AI → Since I have no knowledge of languages other than Japanese and English, translations may contain strange translations. If you notice any translations that need correction, I would appreciate it if you could send feedback on what to change and how! Version 1.2.0 is the biggest update to date and includes numerous new features and bug fixes. Please refer to the release notes for a complete list of changes in version 1.2.0. Release Notes: https://mocolamma.taikun.design/release-notes

Bug Fixes and Improvements - Fix issue where insecure HTTP connections were being blocked → This has resolved the issue where connections to remote servers using VPNs, etc., were being blocked. - Fix padding in the model inspector → An issue where the top and bottom padding was too tight on platforms other than visionOS has been fixed. - Fix issue where model lists could not be sorted in macOS Highlights of New Features in Version 1.2.0 - Revamping the Markdown rendering engine → This incorporating chat improvements such as enhanced performance and syntax highlighting for code blocks, resulting in a more user-friendly experience. → As a result, Mocolamma now requires iOS / iPadOS 18.0 or later, macOS 15.0 or later, visionOS 2.0 or later. Please note that users on older operating systems will need to perform a software update. - Native support for visionOS - Support for attaching images in chat - Support for Ollama's image generation feature (beta) → As Ollama's image generation feature is currently in beta, future Ollama updates may introduce changes to its functionality (Ollama API), potentially causing the image generation feature to stop working in Mocolamma. - Add Image Preview feature → You can preview attached images or generated images in a large enlarged view. - Add an icon customization feature to the add/edit server screen - Add an auto-follow scroll feature to the chat screen and image generation screen - Add many advanced options to custom settings for chat - Add Arabic (ar), Chinese (Hong Kong) (zh-HK), Chinese (Simplified) (zh-Hans), Chinese (Traditional) (zh-Hant), Korean (ko), Russian (ru), Ukrainian (uk) localization by generative AI → Since I have no knowledge of languages other than Japanese and English, translations may contain strange translations. If you notice any translations that need correction, I would appreciate it if you could send feedback on what to change and how! This version is the biggest update to date and includes numerous new features and bug fixes. Please refer to the release notes for a complete list of changes in this version. Release Notes: https://mocolamma.taikun.design/release-notes

Highlights of New Features - Revamping the Markdown rendering engine → This incorporating chat improvements such as enhanced performance and syntax highlighting for code blocks, resulting in a more user-friendly experience. → As a result, Mocolamma now requires iOS / iPadOS 18.0 or later, macOS 15.0 or later, visionOS 2.0 or later. Please note that users on older operating systems will need to perform a software update. - Native support for visionOS - Support for attaching images in chat - Support for Ollama's image generation feature (beta) → As Ollama's image generation feature is currently in beta, future Ollama updates may introduce changes to its functionality (Ollama API), potentially causing the image generation feature to stop working in Mocolamma. - Add Image Preview feature → You can preview attached images or generated images in a large enlarged view. - Add an icon customization feature to the add/edit server screen - Add an auto-follow scroll feature to the chat screen and image generation screen - Add many advanced options to custom settings for chat - Add Arabic (ar), Chinese (Hong Kong) (zh-HK), Chinese (Simplified) (zh-Hans), Chinese (Traditional) (zh-Hant), Korean (ko), Russian (ru), Ukrainian (uk) localization by generative AI → Since I have no knowledge of languages other than Japanese and English, translations may contain strange translations. If you notice any translations that need correction, I would appreciate it if you could send feedback on what to change and how! This version is the biggest update to date and includes numerous new features and bug fixes. Please refer to the release notes for a complete list of changes in this version. Release Notes: https://mocolamma.taikun.design/release-notes

New Features - Add German (de), Spanish (es) localization by generative AI → Since I have no knowledge of languages other than Japanese and English, translations may contain strange translations. If you notice any translations that need correction, I would appreciate it if you could send feedback on what to change and how! - Add access to each tab in the “View” menu → Added menu items to the “View” menu in the menu bar to open the “Server,” “Model,” and “Chat” tabs. Press ⌘ (Command) + 1 to open the “Server” tab, ⌘ (Command) + 2 for the ‘Model’ tab, and ⌘ (Command) + 3 for the “Chat” tab. - Add demo mode → I implemented a demo mode that allows testing the app's basic functionality. → You can access demo mode by adding a server with hostname "demo-mode" (server name is optional) and selecting it. Bug Fixes and Improvements - Fix issue that scrollEdgeEffect did not work in chat view → On iOS / iPadOS 26 and macOS 26 and later, the chat input field background now blurs to blend in. - Fix issue where network requests could fail when launching the app → Since this issue occasionally occurred in the previous fix, I added a retry function to the network check performed when the app launches to ensure this issue is resolved. - Fix issue where the copy button and retry button icons were not displayed in the chat screen on iOS / iPadOS 17 and macOS 14 - Fix issue where the stop button could not be pressed when “Select Model” was chosen in the model picker of the chat screen

New Features - Add French localization by generative AI → Since I have no knowledge of French, the translations may contain strange translations. If you notice any translations that need correction, I would appreciate it if you could send feedback on what to change and how! Bug Fixes and Improvements - Fix “Click” to “Click or tap” - Fix issue where network requests could fail when launching the app

Bug Fixes and Improvements - Fix issue where the icon size for the about this app screen was incorrect on macOS Sequoia or earlier - Fix issue where the model list did not display correctly when models with the same digest as the latest tag existed simultaneously - Fix issue that text conversion couldn't be confirmed while typing in languages requiring text conversion within the message edit field on the chat screen - Improve chat screen to automatically scroll after sending a message - Improve the size of the chat send button on iOS / iPadOS versions