Google appears to be developing a new “Superprojects” system for Gemini that could turn the AI app’s existing project and notebook capabilities into more persistent, goal-oriented workspaces. Evidence found in recent Android app teardowns points to an internal project system called “Robin Superproject,” with features for organizing sources, asking questions, scheduling tasks and setting custom instructions.
The feature has not been officially announced by Google, and its final name, capabilities and release timing remain uncertain. However, multiple rounds of Android app analysis have uncovered increasingly detailed pieces of the system, including a project dashboard, category-based organization and 19 project icons covering areas such as business, coding, education, finance, research, travel and writing.
Key takeaways
- Google is reportedly developing Gemini “Superprojects,” internally referred to as Robin Superproject.
- The feature appears to be an evolution of Gemini’s existing project/notebook concept.
- App resources indicate users could add sources, ask questions, schedule tasks and set project-specific instructions.
- Recent teardown evidence identified 19 project categories/icons.
- Categories include Business, Code, Creative, Education, Finance, Fitness, Research, Shopping, Sports, Travel and Writing.
- The system appears designed around persistent goals rather than isolated chatbot conversations.
- Google already has official Gemini Notebook and Drive Projects capabilities that provide some of the foundation for this direction.
- Google has not confirmed that the leaked “Superprojects” design will launch in its current form.
What are Gemini Superprojects?
The reported Superprojects feature appears to be Google’s attempt to give Gemini a more persistent workspace for long-running goals.
Instead of opening a new conversation every time a user wants help with a particular subject, a project could maintain the relevant context in one place.
For example, someone planning a trip could create a travel project and keep research, conversations, documents and instructions together. Someone starting a business could create a business project containing market research, planning documents and previous discussions.
The same concept could apply to coding, education, finance, writing, fitness or research.
The evidence comes from analysis of Google’s Android app resources rather than an official product announcement. Jetstream reported in September that analysis of Google app version 17.56.14 revealed resource strings associated with a Gemini feature called Projects, with the codename Robin Superproject.
Those resources included interface elements for adding sources, asking questions, scheduling tasks and setting project instructions.
That suggests the project is intended to do more than simply organize conversations.
Google is building on Gemini Notebook
Superprojects would not represent Google’s first attempt to give Gemini persistent context.
In April 2026, Google introduced Notebooks in Gemini, describing them as dedicated spaces for organizing chats and files around complex projects. Google said notebooks could hold conversations, custom instructions and relevant files while synchronizing with NotebookLM.
The original system was designed around exactly the problem Superprojects appears to be addressing: keeping information together when a task takes longer than a single AI conversation.
Google described notebooks as personal knowledge bases shared across its products. Users could move previous chats into a notebook, add documents and PDFs, provide custom instructions and use Gemini against the selected sources.
The company subsequently expanded the concept.
In July, Google renamed NotebookLM to Gemini Notebook, while maintaining synchronization between the standalone research product and the Gemini app. Google said users could access and create notebooks directly inside Gemini and that the notebooks would increasingly work across Google’s ecosystem.
That progression provides important context for the Superprojects leak.
Rather than introducing an entirely separate idea, Google appears to be experimenting with another layer on top of its existing persistent-context system.
The reported project dashboard could do four things
One of the most revealing pieces of evidence came from resource strings discovered in Google’s Android application.
Jetstream’s September analysis found references to four major actions inside a project:
| Reported Project Function | What it could mean |
|---|---|
| Add sources | Attach files or other information as project context |
| Ask a question | Continue a project-specific Gemini conversation |
| Schedule a task | Create a future action related to the project |
| Set project instructions | Tell Gemini how to behave within that workspace |
The fourth capability is particularly important.
Normal chatbot conversations are usually governed by instructions contained within an individual prompt or conversation. A project-level instruction system could allow users to establish persistent rules.
For example, a user might tell a business project to always use a particular writing style, ask a research project to prioritize supplied sources or tell a coding project to follow a particular programming framework.
Those instructions could then remain associated with the project instead of being repeatedly pasted into new chats.
“Schedule a task” hints at a more agent-like Gemini
The reported “Schedule a task” option could be one of the most significant aspects of the project system.
Google has increasingly been moving Gemini from a question-and-answer chatbot toward an assistant that can perform actions and work across applications.
A persistent project with scheduled tasks would fit that broader direction.
Consider a research project.
Instead of asking Gemini once to summarize a group of documents, a user could potentially keep those sources attached to a project and schedule a recurring task around them.
Similarly, a business project could potentially retain instructions and source material while supporting recurring work.
However, this is where it is important to separate what has actually been discovered from what remains speculation.
The Android resource strings demonstrate that Google has been preparing an interface containing a task-scheduling function. They do not establish exactly what scheduled tasks will be capable of doing when or if the feature launches.
Google could change or remove the functionality before a public release.
Gemini Projects could cover 19 categories
A later analysis of Google app version 17.61.19, released September 24, provided more evidence about the direction of the feature.
Jetstream found 19 project icon/category resources associated with Robin Superproject.
They include:
- Business
- Code
- Creative
- Education
- Event
- Finance
- Fitness
- Folder
- Food
- Health
- Home
- Music
- Pets
- Research
- Shopping
- Sports
- Star
- Travel
- Writing
The breadth of the categories is notable.
They are not limited to traditional productivity activities such as work and research. The list also covers personal activities such as food, fitness, pets, travel and shopping.
That suggests Google could be designing Projects as a general-purpose organizational layer for Gemini rather than as a professional productivity feature alone.
A user might therefore have multiple persistent Gemini environments, each dedicated to a different objective.
From chatbot conversations to persistent goals
The bigger change is conceptual.
Most AI chat interfaces are organized around conversations.
A user asks a question, receives an answer and eventually starts another conversation. Even when the system has memory, the interface itself remains centered on individual interactions.
Projects reverse that structure.
The project becomes the primary object, while conversations become one component inside it.
That distinction matters because many real-world tasks do not end after one question.
Planning a vacation can take weeks. Building a business plan can take months. Learning a programming language can take months or years. Researching a complicated subject may involve dozens of documents and hundreds of questions.
A persistent project gives an AI system somewhere to store the context associated with those activities.
Google’s existing Notebook system already moves in this direction.
Google’s official Gemini help documentation says notebooks can provide a continuous chat experience in which Gemini remembers the notebook’s sources and instructions. Google also lists use cases such as planning trips, applying for jobs, learning skills, creating business plans and preparing for exams.
Superprojects could potentially extend this model beyond notebooks into a broader project-management environment.
Google is also connecting Gemini with Drive projects
There is another important piece of the puzzle: Google Drive Projects.
Google’s official Drive documentation now describes Projects as a way to lock a specific set of source files, folders or emails into a project so users can have multiple conversations using the same foundation.
Drive Projects can also be referenced from Gemini side panels across Google Workspace applications, including Docs, Sheets, Slides and Gmail.
That means Google is already building infrastructure in which an AI project can persist beyond a single chatbot window.
This makes the reported Superprojects development more significant.
If Gemini projects eventually connect the app’s conversations, files, Workspace information and task capabilities, Google could create a common organizational layer across its AI products.
The relationship with Gemini Notebook is still unclear
One of the biggest unanswered questions is what happens to the existing Gemini Notebook system.
Current Google documentation says notebooks are already designed to organize projects in a dedicated space, maintain ongoing conversations and remember sources and instructions.
Google also has Gemini Notebook as a standalone research environment.
The reported Superprojects system therefore appears to overlap with capabilities Google already offers.
There are several possible explanations.
Superprojects could eventually replace or absorb part of the current project system. It could become a broader layer that sits above Gemini Notebook. Alternatively, Google could be testing different organizational systems simultaneously and ultimately ship only some of the features.
At this point, there is not enough official information to determine which direction Google will choose.
The “Robin Superproject” name is an internal clue, not a product name
The term Robin Superproject deserves some caution.
The phrase appears in resource names uncovered through app analysis, making it useful evidence that Google is developing something internally under that identifier.
But internal codenames frequently change before products reach consumers.
The final public-facing feature could be called Projects, Superprojects, Gemini Projects, or something else entirely.
Therefore, “Superprojects” should currently be treated as a reported development name, not as a confirmed commercial product name.
Why Google may want a project-based Gemini
The shift toward persistent projects fits a broader change in the AI industry.
The first generation of consumer AI assistants largely competed on answer quality.
The next generation is increasingly competing on how much work an assistant can complete.
That requires context.
An AI agent working on a complex task needs access to relevant files, previous decisions, instructions, tools and sometimes a schedule.
A project can act as the container for those components.
For Google, this approach also has an ecosystem advantage.
Google already controls Gmail, Drive, Docs, Sheets, Calendar, Search and Android. If Gemini projects can connect information across those services while retaining a user’s project-specific context, Google has the infrastructure to build a deeply integrated AI workspace.
That is potentially more powerful than a standalone chatbot.
Google could turn Gemini into a long-term work partner
Imagine creating a project called “Launch My Business.”
The project could contain market research stored in Drive, previous Gemini discussions, financial spreadsheets, a business plan, instructions about the company’s strategy and scheduled tasks.
The user would not need to repeatedly explain the background.
Gemini could theoretically operate against the same body of information every time the project is opened.
A similar structure could work for a coding project.
A developer could keep requirements, documentation, code-related conversations and project instructions together. A writing project could maintain research, drafts and editorial guidelines. A study project could contain notes, quizzes and progress information.
These examples are partly illustrative rather than confirmed Superprojects capabilities, but they demonstrate why Google’s reported design is strategically important.
Google has already tested goal-oriented AI experiences
Google’s current Gemini Notebook products provide another indication of the company’s direction.
Its study notebooks, for example, go beyond storing documents. Google says they can assess a student’s knowledge, generate personalized lessons, track progress and recommend what to study next.
That is already closer to a goal-oriented assistant than a conventional chatbot.
Google has also been expanding Gemini’s ability to work across its ecosystem and introducing more agent-like capabilities.
Superprojects could provide the organizational structure required for those capabilities to operate over longer periods.
Instead of an AI answering isolated questions, the project becomes a continuing relationship between the user, the AI and a defined objective.
The biggest opportunity is context management
The real competitive advantage may not be the project interface itself.
It may be context management.
As AI systems become more capable, users increasingly have more information to give them.
A project might contain documents, emails, notes, previous chats, research, spreadsheets, images and instructions.
The challenge becomes determining which information is relevant at the right moment.
Google has an unusual advantage here because much of that information already lives inside its ecosystem.
A successful Superprojects system could make Gemini more useful precisely because it can connect these existing sources without forcing users to manually assemble the context every time.
Privacy will become more important
A deeper project system also creates new privacy considerations.
If a project can combine conversations, files, emails, calendar information and other Workspace content, users need clear controls over what Gemini can access and how that information is used.
Google’s current Drive Projects documentation already describes source selection and project-level sharing controls.
Any future Superprojects system would need similarly clear boundaries, particularly if it expands into scheduled tasks and more autonomous behavior.
The more persistent the AI becomes, the more important it is for users to understand exactly what context the system remembers and which sources it can access.
What Google has confirmed — and what it has not
At present, the evidence can be divided into two categories.
Confirmed by Google: Gemini has notebooks designed for persistent projects, chats, sources and instructions. Google also has Drive Projects and is connecting notebooks across Gemini and its broader ecosystem.
Reported through app analysis: Google is developing a Gemini Projects system internally associated with the name Robin Superproject, including project categories, custom instructions, source management and task scheduling.
Not confirmed: A public launch date, final product name, final feature set or whether Superprojects will completely replace Gemini Notebook projects.
That distinction is important because app teardown discoveries represent software under development, not a guarantee of what users will eventually receive.
The Bigger Picture
Google’s reported Superprojects work points toward a broader evolution of Gemini from a chatbot into a persistent AI workspace.
The underlying idea is simple: instead of asking Gemini to start from zero every time, users create an environment where the AI already understands the goal, has access to relevant sources and follows predefined instructions.
Google has already established many of the pieces through Gemini Notebook, NotebookLM, Drive Projects and Workspace integrations.
Superprojects could become the layer that brings those pieces together.
If Google succeeds, the result could make Gemini less like a search box and more like a long-term digital work partner. The competition would then shift from who provides the best single AI answer to who can maintain the most useful context while helping users accomplish complicated goals.
Looking Ahead
The immediate question is whether Google will turn the Robin Superproject experiments into a public Gemini feature. The Android evidence suggests active development, but Google has not announced a final product or launch schedule. More app updates should reveal whether the project dashboard, 19 categories and task-scheduling functions remain part of the final design.
If the system launches broadly, its importance will depend on how deeply Google connects it with Gmail, Drive, Calendar, Docs, Sheets, Gemini Notebook and other services. The most compelling version would not simply organize chats; it would give Gemini a persistent workspace in which context, sources, instructions and tasks stay connected over time.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
