Welcome to version 2.5.1 of Perspective Intelligence! Here’s what’s new: • More reliable automatic conversation titles when using private, on-device models. • A more stable and accessible model picker with consistent results while browsing. • Improved OpenAI-compatible provider support, including smarter chat-model selection and clearer capability information. • More reliable app-tool handling for cloud models, with graceful recovery when tool instructions exceed a model’s context limits. • Improved attachment recovery, model availability checks, and overall stability.

Welcome to version 2.5 of Perspective Intelligence! Here's what's new in this big release: • A redesigned model picker makes it easier to choose between Apple Intelligence, on-device models, Ollama, and connected cloud providers. • Expanded private, on-device AI options on iPhone and iPad, with improved LiteRT and MLX support and Thinking controls for supported models. • Smarter multi-step tools, improved memory search, and more relevant personalization using your Identity profile. • Better document grounding for PDFs, Office files, and spreadsheets, along with improved paste and drag-and-drop support. • Improved vision-model selection and more reliable follow-up questions about previously shared images. • Enhanced support for OpenAI, Anthropic, Gemini, and compatible cloud providers. • New daily tips help you discover useful features and can be disabled anytime from Help. • More reliable longer conversations, clearer error messages, and automatic recovery from common model and context-limit issues. • All Access subscribers now enjoy an ad-free experience.

Welcome to version 2.4 of Perspective Intelligence! Here's what's new in this version of the app: • Perspective Intelligence now has an even better agentic harness. The app uses Foundation Models and your favorite models to solve real world problems and work with tools. • Fixes to many tools including Calendar, Web Search, Image Playgrounds and more.

Welcome to version 2.3 of Perspective Intelligence! This version of the app works to improve the initial model selection for initial setup, so you get the best results right out of the box. We've also made improvements to the harness so agentic experiences work even better.

Welcome to the latest version of Perspective Intelligence! This version is mainly a bug fix version, but much of the vision model selection areas have changed, so it gets its own version.

Welcome to version 2.1 of Perspective Intelligence! Here's what's new in this version: • This version of the app now includes the ability to use voice prompts with video mode. Tap the voice button to change the prompt for video mode. This feature requires an all access subscription for Perspective Intelligence • This version fixes an issue where users could not download models for vision modes while using VoiceOver.

Welcome to Perspective Intelligence 2.0, the next era in on-device AI. Here's what's new in this version. • Hybrid on-device chat pairs Apple Intelligence with Gemma 4 E2B for longer local responses. • Expanded attachment support adds PDFs, text files, Markdown, CSV, JSON, HTML, and OCR-backed extraction. • New File Writer support lets Perspective create and update shareable files directly in chat. • Personalization is better with identity setup, contact import, and improved onboarding. • Reminders and Daily Brief are smarter, with structured reminder checklists and clearer source handling. • Vision gets live-video update controls plus improved light detector audio and haptic feedback.

Welcome to the latest version of Perspective Intelligence! Here's what's new in this version of the app: • Daily Brief - You can now receive a daily brief if you are an all access user. This allows you to receive an AI generated summary of everything added to your RSS feeds, Calendars, Reminders and more! • Various AI and Accessibility bug fixes.

Welcome to the latest version of Perspective Intelligence! Here's what's changed in this version of the app: • Light detection in Vision - This version of the app includes a light detection feature that plays tones as the light changes in the room. • Various accessibility bug fixes throughout the app • Gemma 4 12B for many Mac users

Welcome to version 1.15 of Perspective Intelligence! Here's what's new in this version of the app: • New Knowledge support for Apple Intelligence conversations, starting with an Accessibility Guide adapter for WCAG 2.2 AA guidance. • A redesigned chat model picker makes it easier to choose Apple Intelligence, on-device models, Ollama, OpenAI-compatible providers, and vision models. • Tools and Agents now live in the message add menu, keeping the composer cleaner while making advanced workflows easier to reach. • Vision mode has been refreshed with separate Photo and Video workflows, model selection for live video, description detail settings, and better follow-up handling. • Image conversations are smarter, with improved routing for attached images and follow-up questions about previously shared images. • Responses should be less likely to stop early thanks to improved response budgeting across Apple Intelligence, Gemma, Ollama, and OpenAI-compatible providers. • Added a new Help section with guidance for chat, tools, agents, models, attachments, voice, and vision features.

Welcome to version 1.14 of Perspective Intelligence! Here's what's new in this version of the app. • Added support for connecting to Hermes Agent • We've added full support for OpenAI compatible AI providers with presets. This replaces the specific oMLX implementation. Don't worry, oMLX is a preset. • All access members can now add custom prompts in video mode. Now, you can create your own video mode. • You can now select a downloaded model from the Suggested Prompts/Conversation Tools dialog. • Web Search, Deep Research and Page Summarization tools have been switched over to Exa Search instead of Brave Search. • This bug fix version improves OpenAI compatible connection setup. We hope you are still enjoying the app and please send us feedback if you have questions.

Welcome to version 1.14 of Perspective Intelligence! Here's what's new in this version of the app. • Added support for connecting to Hermes Agent • We've added full support for OpenAI compatible AI providers with presets. This replaces the specific oMLX implementation. Don't worry, oMLX is a preset. • All access members can now add custom prompts in video mode. Now, you can create your own video mode. • You can now select a downloaded model from the Suggested Prompts/Conversation Tools dialog. • Web Search, Deep Research and Page Summarization tools have been switched over to Exa Search instead of Brave Search. • This version now fixes an issue where users could not edit or delete custom video mode prompts. We hope you are still enjoying the app and please send us feedback if you have questions.

Welcome to version 1.14 of Perspective Intelligence! Here's what's new in this version of the app. • Added support for connecting to Hermes Agent • We've added full support for OpenAI compatible AI providers with presets. This replaces the specific oMLX implementation. Don't worry, oMLX is a preset. • All access members can now add custom prompts in video mode. Now, you can create your own video mode. • You can now select a downloaded model from the Suggested Prompts/Conversation Tools dialog. • Web Search, Deep Research and Page Summarization tools have been switched over to Exa Search instead of Brave Search. • This also fixes a bug on the models grid where the app can slow down while picking a model. We hope you are still enjoying the app and please send us feedback if you have questions.

Welcome to version 1.14 of Perspective Intelligence! Here's what's new in this version of the app. • Added support for connecting to Hermes Agent • We've added full support for OpenAI compatible AI providers with presets. This replaces the specific oMLX implementation. Don't worry, oMLX is a preset. • All access members can now add custom prompts in video mode. Now, you can create your own video mode. • You can now select a downloaded model from the Suggested Prompts/Conversation Tools dialog. • Web Search, Deep Research and Page Summarization tools have been switched over to Exa Search instead of Brave Search We hope you are still enjoying the app and please send us feedback if you have questions.

This version of Perspective Intelligence fixes an issue that causes the chat input field to not be activated while VoiceOver is enabled.

Welcome to the latest version of Perspective Intelligence! Here's what's new in this version of the app. • High quality Text To Speech voices are now available throughout Perspective Intelligence • You can now search through all models that will work on your device from mlx-community. Head to Settings, and then AI Models to make a selection. • The contacts tool can now load more than one contact. Just ask, "Search for all contacts named Michael" in a chat. This will require All Access • All Access members can now use the Speak tool. use @speak to create a spoken audio file with the specified voice. We hope you are enjoying Perspective Intelligence. Please send any questions or feedback to feedback@perspectiveintelligence.app

This version of Perspective Intelligence fixes a crash when pasting images for description, and an issue where the app would simply return the image resolution instead of a description while in chat.

This version of Perspective Intelligence resolves an issue that users have found where models could not be downloaded in Settings while using VoiceOver.

Welcome to version 1.11 of Perspective Intelligence! Here's what's new in this version of the app. • Video Mode - You can now use our new live video mode to describe what's happening in the world around you. Just activate the video button in the Vision mode. • Improved Gemma 4 support - Perspective Intelligence improves on its support for the Google Gemma 4 models. • Improved Tool Routing - Perspective Intelligence is now smarter with better tool routing. There is also an option to turn tool routing off and just try out the model without the possibility of using tools. • LFM 2.5 450M for Vision - We've added the amazing Liquid AI LFM 2.5 450M Video Model for use with Perspective Intelligence. You can use this model for video mode, or for describing images. This model is small and extremely fast while providing the best accuracy. We hope you are enjoying Perspective Intelligence, and we welcome any feedback you might have to make the app even better.

Welcome to version 1.10 of Perspective Intelligence! Here's the exciting new features in this version. • We've added Google's latest model to the app. Perspective Intelligence now includes Google Gemma 4 as a model in the app. This model is still experimental, and may have inconsistencies compared to other models. • Deep Research Agent is now here. You can now have Perspective Intelligence conduct research on the web for you. This agent is an All Access agent, and will show what sources it has used to provide information. • You can now summarize web content with Perspective Intelligence's Page summary tool agent. Just give the agent a url and it will summarize a page for you using your selected model. We hope you enjoy this new version of the app, and please give us feedback on the use of Gemma 4 as you use the app. Please note that not all devices will support Gemma 4 E2B. This model is still experimental, so please keep this in mind as you select it.

Welcome to version 1.8 of Perspective Intelligence! This version of the app includes the following features: • New shortcuts actions! You can now say to Siri, "Ask Guide in Intelligence" to ask the Perspective Intelligence App Guide a question. This works with any other built in tool or user agent. • Menu bar support for the mac version of the app. You can control Perspective Intelligence from the menu bar of the mac version of the app. We hope you are enjoying the app, and please let us know in the app if we can make improvements in any way.

This version of the app resolves an issue where the memories tool agent did not work correctly for all users.

Welcome to version 1.7 of Perspective Intelligence! Here's what's new: • Added device tool agent for all users. You can now ask questions about your device and Perspective Intelligence can assist with questions like your battery level, screen brightness and more! • All Access members can now search the web when a general response isn't very accurate. Just use the Search the web instead button after a responsee.

Welcome to 1.6 of Perspective Intelligence! Here's what's changed: • New chat and vision models to support certain iPhones and iPads. • Live audio transcription, and memories are now available to all users. You will no longer need to be an All Access member. Transcribing files will still require All Access. • Users can now try All Access for one week. Check this out in the app. • There are now fixes for certain VoiceOver bugs that existed in version 1.5 for vision users. • Tools are now Tool Agents. ask a tool agent to complete a task, or let the app route your request to one of these agents. • All Access members can create custom persona user agents with their own abilities. These include temperature, model, instructions, and more. All users can use certain built in personas which include the app guide, coder agent, and writing agent.

Welcome to Perspective Intelligence 1.5!  We are so excited that so many users have tried the app. Here are the improvements that we've made this version: • Improved chat experience - The output from the AI now contains rich content and better flow to be more natural, and VoiceOver users can now interact with each element of the AI's responses. It can even do tables! • Audio transcription from files - You can now use Whisper to transcribe audio from files This is an all access feature since it is audio transcription • Improved model selection - Downloading models works much better in this version, and we recomend people use smaller models to be more efficient. Also, you now have the choice to pick the model you like from a list. • Improved vision features - We've refined how the app detects and reads text and other elements in the vision section of the app. • Compose Email - Use AI to create an email. This will open the email composer to send your email. This is an All Access feature. We always value your feedback, and we hope you continue to help us improve the experience of Perspective Intelligence