ChtSafe 2.7.4 is all about polish — dozens of fixes from our latest quality review. • Read aloud works again, and saved audio files contain the real recording • Dictation now transcribes your voice reliably • Pictures from your library open properly instead of an empty frame • Regenerated answers appear right away, and alternatives are kept when you return to a chat • Select and copy just part of a reply • iPad: signing in no longer freezes after using a password manager, and settings pages open beside the sidebar • A Forgot Password? link on the sign-in screen • Picking a model is easier: the list is searchable, Auto sits on top, and "Help me choose" suggests one for you • The keyboard no longer covers the send button in the full-screen editor • Report a Bug and Request a Feature now take screenshots and confirm what you sent • Release notes like these now live in Help & Docs Thanks for all the reports that shaped this release — keep them coming via Report a Bug in Settings.

ChtSafe 2.7.3 is a complete rebuild — faster, calmer, and more reliable from the first tap. • Streaming chat with visible thinking and tool steps • @-mention knowledgebase files and connected folders (Nextcloud, OneDrive, SharePoint) • Generate images, music, video, and audio right in the chat — then refine with follow-up prompts • Cross-device continuity: start on your desktop, pick up on your phone • Automatic recovery from network drops • New desktop apps for macOS, Windows, and Linux (beta)

- Added Nano Banana 2 image generation through Gemini 3.1 Flash Image. - Expanded the model lineup with GPT-5.4, Claude 4.6, Gemini 3.1 Pro, and new Qwen Next variants. - Updated the default model to GPT-5.4 Enterprise. - Refined Multi-Model presets for stronger comparisons and better result quality. - Included translation fixes plus stability and security updates.

Mostly bug fixes, also we did an optimization for using the app in landscape mode.

Meet the new ‘Secure Media Generation’ Suite, create Songs, Videos and Image with our secure backend, keeping your ideas and data save.

What's new - You can now Flag messages and images generated by AI and submit them for review Bugfix - If a request was solely a research task the action buttons such as "Copy" where not available

What's new - A new onboarding experience for first-time users - Grounding/Deep research now shows progress and sources again (Bugfix and improvement) Bugfix - Plugins selected as default are now properly loaded on startup

What's new Fully reworked Signup and Signin experience, you can now signup with just 1-Click either via Mail or Google, Apple, Microsoft, LinkedIn or X. What's fixed - Lots of navigation errors when using the App - Smaller annoyances in the image generation

What's new - Agents are now fully supported on Mobile - Secure image processing is now fully supported on Mobile - Secure Media Generation has been added with Google Gemini, xAI Aurora and OpenAI Image-GPT1 What's fixed - Both the Sign-Up and Sign-In have been completely redone to be easier and more reliable - Full language Support for Japanese, German, English, Spanish, Russian and Ukrainian as well as many more More details - You can now submit and manage both Bug reports and Feature requests

What's new - You can now use long-tap to copy and edit your own messages - The settings menu has been re-organized What's fixed - a couple of dark mode display issues - markdown rendering did not always work

This is a minor bugfix release, resolving an issue around inApp purchases. We had a race-condition which in around 30% of the cases prevented users from immediately receiving their purchase confirmation.