VisionFeed 3.0 brings numerous new features and improvements: - Significantly improved Reader view with more reliable article detection and image display - Optimised workflow for translating articles and titles - AI chats now appear in a clearer, dedicated view - New feed rules: pause updates and automatically mark new articles as read - New status overview for faulty, paused and unreviewed feeds - Improved display of iCloud sync status - Additional filters for translated articles, articles without images and specific time periods - Updated AI models, cost calculation and web search - Completely revised translations for all 44 supported languages - Numerous improvements to usability, appearance, stability and performance

This update overhauls the AI features and improves usability: • AI conversations now open in a clear, dedicated view • Updated AI models and cost calculations • Improved web and feed search • Redesigned AI settings and API key management • Extensive translation corrections and improvements • Further bug fixes and minor improvements

AI features significantly improved • AI chats now open in a dedicated, clearer view. • Support for GPT-5.6 Terra (Mini) and GPT-5.6 Luna (Nano). • Model selection, recommendations, and cost estimations have been updated. • Web search and automatic RSS feed discovery now deliver more up-to-date and reliable results. • Improved management of personal OpenAI API keys. • Numerous fixes for reader view, feed checking, and article analyses. • Translations and user interface have been thoroughly revised.

The AI features of VisionFeed have been thoroughly modernised. Users can now enter their own OpenAI API key and pay for usage directly through their OpenAI account. The previous credit packages and in-app purchases have been removed. New additions include a detailed cost overview, new GPT models, improved error messages, and numerous optimisations for iOS 26 and Swift 6.

VisionFeed has been updated to iOS 26 (iOS 26 is now required). We have also modernised and simplified the AI backend: Gemini and GPT-4.1 have been removed, and VisionFeed now uses GPT-5 Nano and GPT-5 Mini. For article translation, there is a new setting for quality control (Reasoning Low/Standard), while title translation remains cost-optimised. In addition, several iOS 26-specific UI issues have been fixed to ensure that buttons and actions in portrait mode work reliably again.

What's new in this version • Feed sharing – RSS feeds can now be easily shared with others. • Improved icon loading – Feed favicons are now detected and displayed more reliably. • Manual icon reloading – Using the new context menu, a feed's icon can be specifically updated as needed. • Optimised OPML import – Importing feeds from OPML files has been improved and is now more stable. • New setting for parallel updates – You can now specify in the settings how many feeds are loaded simultaneously during refresh. • Better handling of article images – Images in articles are now loaded and displayed more reliably. • General bug fixes – Numerous minor issues have been resolved and overall stability has been further improved.

What's new in this version • Switch to GPT-5 – VisionFeed now specifically utilises the most efficient GPT-5 models for even better results at significantly lower costs. • Massive cost reduction – By using GPT-5 Mini and, above all, the performance of GPT-5 Nano, costs have noticeably decreased across all areas. • Optimised reasoning – Every task now uses tailored settings for the perfect balance between quality, speed and cost. • Stability & performance optimisations – Improved processing for translations and AI-powered features for an even smoother user experience.

• The AI feed search has been completely redesigned – faster, more precise, more helpful • Improvements to AI chats, including optimised link handling • The feed checking tool has been expanded and simplified • Improved swipe gesture in the article view – now easier to trigger • Localisation added to the AI settings

• Improved code structure: The code has been restructured and revised, further enhancing the maintainability, stability, and scalability of the app. • Source display for articles: The source is now shown in the article view (including ReaderView), so it is always clear which website the article comes from. • Bug fix in portrait mode: The display issue with cards (CardView) in portrait mode on the iPad has been resolved. • Faster AI translation: Article translation via AI is now significantly faster and more reliable – even for long articles. • AI settings via iCloud: Your AI settings (e.g., model selection) are now automatically synchronised via iCloud – so all your devices always stay up to date.

- VisionFeed is now subscription-free – you only pay for the AI features you actually use. - New powerful AI models: Gemini 2.5, ChatGPT 4o and 4.1 are now supported. - AI model selection can now be made separately for different functions. - Additional settings are automatically and securely synchronised via iCloud. - Articles can now be added to favourites directly from the article view. - Image management and feed management have been comprehensively revised. - There is a new billing page where you can transparently track your AI usage. - The user interface has been modernised and improved. - Many localisations have been corrected and new languages have been added.

What's new in this version: • Modified article titles now correctly appear in the list • Improved feed search: CDATA elements in titles are removed • New help section added to the main screen • Startup notice added for large CoreData databases • Improved sync of the "maximum article age" setting with a default value