Google is advancing the projects feature for its Gemini AI in the Android app, enabling direct creation on mobile devices. This development, spotted by Android Authority, aims to compartmentalize user interactions and files, similar to ChatGPT’s approach. Updates include UI reorganizations for better access.
AI tools such as Gemini demonstrate utility across various tasks, yet users often require methods to maintain focused and separated interactions. ChatGPT addresses this need through projects, which serve as dedicated workspaces that organize conversations and files related to specific work. Gemini has been observed developing a comparable solution. Initial signs appeared earlier this year, indicating Google’s intent to introduce structured project environments within its AI platform.
In September, examination of the Gemini Android app revealed a “Projects” label integrated into the side hamburger menu. Accompanying text suggested that users would need to initiate Gemini projects via a web browser before accessing them on the mobile application. This setup implied a dependency on desktop or web interfaces for project setup, limiting the immediacy of mobile-based workflows.
Recent modifications in Google app version 16.47.49.ve.arm64 mark a shift in this process. The update introduces a “New Project” button directly within the Gemini interface on Android devices. This addition allows users to begin project creation without relying on external web access, enhancing the app’s self-sufficiency for such features.
Although the “New Project” button appears and can be interacted with, the “Create project” option remains non-operational. Tapping it displays the button but does not proceed to functionality, indicating that the feature is still in preparatory stages. Developers have not yet enabled full interaction, which restricts current testing and detailed observation of the project’s mechanics.
A newly identified string in the app code provides insight into the feature’s design. The string states: “Add up to 10 files that Gemini can reference throughout your project.” This element supports the concept of projects as containers for relevant resources, ensuring that Gemini draws from designated files without incorporating them into unrelated tasks. The limit of 10 files aligns with efforts to manage scope and prevent overload in AI processing.
Details on the file limit’s application remain unspecified. It is not confirmed whether the 10-file cap applies universally or adjusts based on the user’s Google AI plan subscription. Different tiers of service might influence storage or reference capacities, but no explicit information clarifies this aspect in the current build.
Alongside projects, the app incorporates preparations for subtle interface adjustments. Currently, the account switcher lacks a direct shortcut to NotebookLM, Google’s note-taking and research tool integrated with AI. Code changes suggest developers plan to include this shortcut, streamlining navigation between Gemini and related services.
Access to Gemini’s scheduled actions presently requires navigation through the account switcher followed by settings. This path positions the feature somewhat deeply within the menu structure. Indications point to a potential relocation of scheduled actions to the side hamburger menu, which would place it more prominently alongside other core options like Projects.
All these enhancements, including the projects button, file reference string, NotebookLM shortcut, and menu reorganizations, exist solely in the development version of the app. They are not accessible to general users at this time. Google retains control over the rollout timeline, determining the precise moment for public deployment of each component.
