Iris is an intelligent assistant application running on macOS, with the goal of working across all of the user's applications, acting as an assistant that understands the user's intent and can proactively fill in information gaps. According to the official introduction, users only need to state their needs, and Iris will complete them across applications, and it will learn user preferences over time, and in the future may even anticipate needs.
-
Purpose and positioning Iris attempts to change the way people collaborate with computers. Officially, traditional applications are isolated from one another, and users need to do a large amount of coordination work themselves, such as copying, pasting, searching, arranging, and clicking menus. Iris is positioned like an excellent assistant: if the user says "schedule a coffee with Sam next week," the assistant will check the calendar, find the contact, choose a nearby cafe, and pick a suitable time. Iris hopes to bring this experience to the computer, letting software be rebuilt around this idea and make decisions in the way users make decisions.
-
Main capabilities
-
LLM client: Iris was originally an LLM client, serving as the foundation of the entire product.
-
Smart calendar: This is the first extension feature added to Iris. Users can create or adjust schedules in natural language, for example "lunch with Tim next week" or "move the dentist appointment to Thursday," and Iris will automatically fill in missing information. At the same time, it retains a complete interface, with all fields and options in their expected positions, combining the speed of natural language with the precision of direct control. The calendar can be used alongside existing calendars or completely replace them. The more it is used, the better Iris can learn the user's habits, such as how long meetings usually last, where they like to have lunch, how much buffer time is needed between calls, and even patterns the user themselves has not noticed.
-
Screenshot tool improvements: The screenshot tool has been improved in the Beta version.
-
Onboarding redesign: Officially, the onboarding was redesigned.
-
Cross-application understanding: Double-click the Option key, and Iris appears at the edge of the screen as a narrow bar. Point it at any application to immediately understand what the user is doing. It floats above windows, has a moderate presence, is not intrusive, and moves with the user. Officially, it is also said that in the future Iris will go deep inside applications and present executable actions directly in the application interface.
-
How to use V. How to use The current version is Beta, and the official invitation is for users to become early testers. The core interactions include: double-click the Option key to bring up the narrow bar at the edge of the screen; point the narrow bar at any application to understand the current work; enter needs in natural language, or use the complete interface for precise operations. Officially, Iris is a Swift-based Mac application.
-
Applicable scenarios Suitable for those who want to reducemanual coordination across applications, quickly complete scheduling through natural language, and let the assistant learn personal work habits for macOS users. It is especially suitable for scenarios involving frequent meeting scheduling and the need to handle the flow of information between multiple applications.
-
Limitations and development status It is currently in the Beta stage, and the official note states that Beta content may change. The calendar is the extension currently released; email, tasks, and notes features are still in development, and the official statement says they will be launched afterward and connected through Iris. The future capabilities mentioned officially include: reading received emails, drafting events before the user speaks, handling scheduling end to end, and going deep inside applications to present actions. These are all planned or future directions and have not all been implemented yet. Information about pricing, cost, and supported platforms (other than macOS) is not yet provided in the materials.
