A Google AI workspace for research, connected sources, editable drafts, notebooks, media, and bounded recurring work.
Conversation & research · Changing · Documentation checked 2026-10-04
Gemini Apps for personal, work, and school accounts; Workspace apps and developer platforms differ.
Age, country, language, device, Google AI plan, and Workspace license affect access.
What the platform is for
Gemini Apps helps people explain topics, investigate evidence, create drafts, and work with selected Google or other service context. A useful starting task has clear input and output: research a public question, summarize an approved file, or refine a short document. Its feature names should be understood as components with different jobs.
Know the product boundary
Personal Gemini, organizational Gemini Apps, Workspace app features, Gemini Notebook, and developer products have different terms and entitlements. Deep Research gathers evidence; Deep Think is experimental reasoning. Scheduled actions prepare responses; Spark is an eligible personal-account agent. A Google account alone does not establish access to all of them.
How the components connect
Files and connected apps supply context. Notebooks organize ongoing sources and discussion. Skills specify a method while memory recalls useful context. Canvas gives a result an editable surface. Live is a supported mobile interaction, and generated images / videos are creative outputs. Each added component changes the source, storage, or action boundary.
A beginner starting point
Begin with public-source research or one approved file. Check sources and calculations, then use Canvas for an editable handout. Explore notebook context before reusable skills, connections, or scheduling. Learn the Gems transition before investing in an older Gem-based setup; the removal timeline differs across personal, work, and education accounts.
Limits and data choices
Plan, age, region, language, device, activity settings, and Workspace license affect access. Enabling activity can unlock several features while changing data handling. Spark has significant regional restrictions. This first edition teaches selected major media and learning surfaces; music, avatars, study variants, and every regional action are not complete walkthroughs here.
The technical path
For developers, use Google’s official Gemini developer and AI Studio documentation rather than assuming the app’s model menu or billing carries over. For organizational use, read the Workspace access and privacy sources first. Evaluation, authentication, deployment, and connector governance require an implementation-specific technical design.
Feature map
Deep Research and source selection
Review an investigation plan and deliberately choose web, uploaded, or connected sources.
Files, analysis, and Canvas
Analyze supplied material and inspect a document, app, or code draft in Canvas.
Skills, Gems migration, and memory
Use task instructions and remembered context while accounting for the current Gems transition.
Connected Apps and action boundaries
Let Gemini use a selected external service with a clear read or action scope.
Gemini Live, camera, and screen context
Use spoken interaction and selectively share visual context on a supported mobile device.
Image and video creation
Create a labeled visual draft while respecting input rights and product-specific access.
Scheduled actions and Gemini Spark
Distinguish recurring assistant requests from an agent that performs ongoing tasks.
Activity, privacy, and Workspace access
Check personal-account data settings separately from organizational data protections.
Notebooks, sources, and ongoing context
Keep a project’s source material and conversations in a connected notebook.
Models, reasoning, and Deep Think
Distinguish an available model from an experimental reasoning level for a difficult problem.
Complete workflows
From research to a Canvas handout
Turn a verified public-source report into a beginner handout and an inspectable prototype.
A notebook and weekly learning review
Keep approved lesson context organized and generate a private recurring review.
Coverage and availability
Models, modes & conversation
Covered: Deep Think experimental reasoning with entitlement; ordinary model availability varies.
Search, research & sources
Covered: Deep Research, source selection, and evidence verification.
Files, analysis & code
Covered: Files, spreadsheet analysis, and Canvas code inspection.
Images, video, audio & voice
Covered: Live, images, and video covered; music and avatars lack full walkthroughs here.
Projects, instructions & memory
Changing: Skills, Gems migration, memory, and connected Notebooks.
Editing surfaces & deliverables
Covered: Canvas docs, apps, slides, and export / public-link boundaries.
Agents, skills, plugins & MCP
Not established: Connected Apps and Spark; arbitrary third-party plugin / MCP app parity not established.
Orchestration & scheduling
Changing: Scheduled actions versus Spark; account and regional restrictions.
APIs, SDKs, deployment & evaluation
Separate product: AI Studio / developer APIs are separate from Apps and Workspace.
Sharing, privacy & governance
Covered: Privacy Hub and work / school terms; activity dependencies explained.
Plans, devices & accessibility
Not established: Age, region, plan, and device conditions included; complete quotas / accessibility behavior not established.
Sources and verification
Documentation checked 2026-10-04. Examples are illustrative unless described otherwise.
Deep Research in Gemini Apps · Google
Upload and analyze files · Google
Create with Canvas · Google
Skills in Gemini · Google
Gems migration to skills · Google
Memory and personalization · Google
Use connected apps · Google
Work and school product boundaries · Google
Gemini Live · Google
Generate and edit images · Google
Generate videos · Google
Scheduled actions · Google
Get started with Gemini Spark · Google
Gemini Apps Privacy Hub · Google
Notebooks in Gemini Apps · Google
Deep Think: experimental reasoning · Google
Gemini API developer documentation · Google