About Gemini
Gemini accepts text and other context to help you write, learn, plan, and create. Deep Research builds a research plan and report; Canvas gives you a workspace for documents and code. Connected Google apps can supply additional context when available and authorized.
Main features
- Ask questions and work with text, images, and audio
- Plan and carry out research with Deep Research
- Draft and refine documents or build prototypes in Canvas
- Generate and edit images, and use Gemini Live
- Connect supported Google apps for relevant context
Inputs, outputs and access
| Item | Details |
|---|---|
| What you can provide | Text, images, audio and files. |
| What you can create | Text, images, voice, code, interactive apps and research reports. |
| Account access | Some web features work while signed out; additional features and saved activity use a Google Account. Work and school access depends on account eligibility. |
| Interface languages | Web languages include Simplified Chinese, Traditional Chinese, English, Japanese, French, German and Spanish. |
| Availability examples | Available in supported markets including the US, UK, Japan, Hong Kong and Taiwan; mainland China access is limited to Workspace. |
Capabilities in practice
Choose the scope of a research report
Deep Research first proposes a plan you can refine. You can add files and choose whether it should use Gmail, Drive and Chat alongside the web. Progress remains visible during research, and the report can later become an Audio Overview.
Turn research into an interactive result
When a Deep Research report finishes in Canvas, its Create menu offers a web page, infographic, quiz or Audio Overview. Canvas also supports focused text edits. Projects can be opened on mobile, while text styling and formatting require the desktop web app.
Platforms and access
Web, Android, iOS, Windows, and macOS; availability and native capabilities vary.
Pricing and plans
US personal-plan monthly prices are USD 4.99 for AI Plus and USD 19.99 for AI Pro. Ultra offers two usage tiers at USD 99.99 and USD 199.99 per month. These prices apply to the United States. API usage is billed separately.
Open source and deployment
The Gemini app and Google Gen AI Python SDK are separate products. The SDK uses the Apache-2.0 licence, which covers the SDK.
Data and privacy
With Keep Activity on, saved activity can help improve Google services and train generative models. Turning it off excludes future chats from model training unless you submit feedback; chats are still retained for 72 hours. Reviewed data can remain for up to three years. Connected apps follow additional policies.
Frequently asked questions
Is Gemini free?
A free personal plan is available with usage limits. Paid Google AI plans increase access and add features.
What can Canvas create?
It can draft documents and generate shareable app or game prototypes from prompts. You can also turn research into quizzes and other interactive outputs.
Is there a desktop app?
Yes. Windows and macOS apps are available; platform requirements and some native capabilities differ.
Does it support Chinese?
The web app supports Simplified and Traditional Chinese. Availability in mainland China is limited to Workspace.
What happens if I turn off Keep Activity?
Future chats are not used to train Google’s AI models unless you submit feedback. They are still saved for 72 hours, and some connected apps become unavailable.
Demos and tutorials
Product updates
- Google rolls out reusable skills in Gemini chat · 2026-09-30
Product and provider
Product scope
Gemini Apps and account-specific Google AI subscription access; Gemini API, Vertex AI and downloadable Gemma weights are separate offerings.
Provider and service
Gemini is provided by Google (Google LLC). The product is available as an online service.
| Tagline | Google’s AI assistant for writing, research, creating content, and working with Google apps. |
| Contact channels | Search the help center for setup, account and billing support. |
Access and help
Official website · Launch url · Pricing url · Documentation url · Support url · Policy url · Terms url
| Legal urls | Service terms, privacy and applicable organization agreements. |
Working style
Interaction and review
Use natural-language conversation to generate, analyze and retrieve information. Editable working surfaces are available for supported deliverables; review the final result before use.
| Primary category | Assistants |
| Product forms | website; mobile app; desktop app |
| Integration ecosystems | Google Workspace, MCP |
Tasks and preparation
People and tasks
The assistant covers general research, writing and file work. It can fit individual, small-team and enterprise workflows, with the relevant plan and organization controls.
| Problems solved | Combine source material, reuse context and prepare a reviewable first draft. |
| Primary jobs | Research a topic and produce a report; Refine a draft or presentation outline; Build a shareable app or interactive learning quiz; Find and summarize information across Google apps |
| Target roles | Students and independent learners; Writers and knowledge workers; People who use Google apps for work |
| Experience level | Basic chat uses natural-language instructions; connected workflows also require permissions and result review. |
| Prerequisites | Prepare the task, relevant files and permission to use them. Sign in for saved or paid features. |
| Value proposition | Reduce manual collection and first-draft assembly while keeping a review step. |
| Unsuitable scenarios | Features and limits depend on account type, age, region, and plan. Canvas text-format editing is available on desktop web, with more limited editing on mobile. The web app supports Simplified and Traditional Chinese; mainland China availability is limited to Workspace. |
| Manual alternatives | The corresponding manual workflow is to read sources, organize findings and edit the final document. |
Files and results
File, video and code limits
A prompt can contain up to 10 files. Most files are limited to 100 MB; videos can reach 2 GB. Total video input is 5 minutes, or 1 hour with AI Pro/Ultra; audio is 10 minutes, or 3 hours with those plans. A code folder or GitHub repository can contain 5,000 files within 100 MB.
From a task to a result
Storage and incomplete analysis
Gemini’s saved-activity storage is separate from Drive storage. For a storage error, delete unneeded activity or Gems and wait for the change to take effect. For partial analysis, divide large files and ask about a narrower range.
Example task: Turn research into a slide outline
Research this fictional product category, show the plan first, and make a five-slide draft in Canvas after I confirm the scope.
Get started
Open Gemini or install an official mobile or desktop app. Ask a question or choose a tool such as Deep Research or Canvas. Add relevant context, review the output, and refine your request.
Example steps
Prepare sample materials you have permission to use. Research this fictional product category, show the plan first, and make a five-slide draft in Canvas after I confirm the scope. Review the draft and save the result.
Platforms and availability
| Support languages | English product documentation is available. |
Subscriptions and billing
| Access model | freemium; subscription |
| Free tier | Free plan available. |
| Free | USD 0 · per account. |
| Google AI Plus | USD 4.99/month · per account. |
| Google AI Pro | USD 19.99/month · per account. |
| Google AI Ultra starting tier | USD 99.99/month · per account. |
| Google AI Ultra higher tier | USD 199.99/month · per account. |
Choosing and using the tool
| Key takeaways | Google’s AI assistant for writing, research, creating content, and working with Google apps. Free; US prices: AI Plus USD 4.99/month, AI Pro USD 19.99/month, Ultra from USD 99.99/month. Features and limits depend on account type, age, region, and plan. |
| Pros | Ask questions and work with text, images, and audio; Plan and carry out research with Deep Research; Draft and refine documents or build prototypes in Canvas |
Apps, integrations and development
Choose apps and account permissions
Connected Apps can use Gmail, Calendar and other approved services. Use @ to select an app, and control connections in Settings. With Keep Activity off, connected apps are unavailable on the web and iOS; Android retains a smaller set of device and communication actions.
Custom MCP apps
Custom MCP apps currently require a personal Google Account, age 18+, a US location, English and Keep Activity enabled. Add the server URL on the web; the connected app can then be used on the web and mobile. Write actions ask for manual confirmation.
Ownership and deployment
Input and output rights
Google does not claim ownership over original content you generate. Material supplied by other people and incorporated third-party content retain their own rights.
Continue editing in Docs, Slides or Colab
Canvas text can open in Google Docs; slides can open in Google Slides or download as PDF. Python code can be exported to Colab. Shared Canvas apps can retain data across sessions, and anyone with a public app link may be able to view and edit its associated data.
Hosted product and generated files
Gemini runs as a hosted product. Installing its official client does not deploy the full service on your own server. Generated files and scripts can be used separately where export is supported.
Move work to other Google tools
Export Canvas slides as PDF, or send supported content to Docs, Slides and Colab. Python code can be exported to Colab for further work.
| Generated code license | Generated code is covered by the output terms; included third-party dependencies keep their separate licences. |
| Third party rights | Third-party material can carry separate terms; a generated response does not itself establish a licence to all quoted material or dependencies. |
Privacy and organization controls
Activity history and retention
With Keep Activity enabled, activity auto-deletes after 18 months by default, with 3-month, 36-month and manual options. Temporary chats and chats with activity off are retained for 72 hours. Human-reviewed content can remain for up to 3 years, disconnected from your account.
Data handled during use
Using the service can involve account details, prompts, files, generated outputs and device or usage information. Paid services also process billing details; optional app connections bring in the information you authorize.
Export and delete activity
Google Takeout exports eligible account data. Delete Gemini activity in Gemini Apps Activity; data in other Google services is managed separately. Reviewed material may be retained without its account link for up to three years.
Working limits
Long sources and shared apps
Large inputs can exceed the working context and leave details out. A public Canvas app link can expose associated data to other users; connected third-party services apply their own data handling.
Export, migration and alternatives
Bring existing chats into Gemini
Eligible personal accounts can import memories or chat-history ZIP files from other assistants. Chat import accepts files up to 5 GB and up to 5 ZIP uploads per day. It requires age 18+ and currently excludes the EEA, Switzerland and UK, as well as work, school and supervised accounts.
A related option
ChatGPT is a partial alternative for research and editable documents. Its working environment and connected-app ecosystem differ. Compare the output you need and required permissions; this is a workflow comparison for 2 October 2026.
Sources and further reading
Information checked: 2026-10-02
Gemini app overview
View source · Google
Topics: Name · Record scope · Tagline · Vendor name · Lifecycle status · Official website · Launch url · Documentation url · Support url · Contact channels · Primary category · Secondary categories
Google AI subscriptions — United States
View source · Google
Topics: Pricing url · Access model · Summary · Free tier · Checked at · Output modalities
Gemini Deep Research
View source · Google
Topics: Example prompts · Use cases · Problems solved · Value proposition · Manual alternatives
Gemini for desktop
View source · Google
Where you can use the Gemini web app
View source · Google
Topics: Interface languages · Offered regions
What you need to sign in to Gemini Apps
View source · Google
Topics: Account requirement · Authentication methods
Google Gen AI Python SDK
View source · Google
Gemini release notes: skills
View source · Google
Google Terms of Service
View source · Google
Topics: Terms url · Legal urls · Vendor legal name · Product open source status · Self host status · Ownership terms · Third party rights · Generated code license
Gemini Apps Privacy Hub
View source · Google
Topics: Policy url · Retention policy · Data categories · Data export process · Deletion process
Upload and analyze files in Gemini Apps
View source · Google
Topics: Input modalities · Limits · Troubleshooting · Known limitations
Create docs, apps and more with Canvas
View source · Google
Topics: Export scope formats · Export formats · Generated code export · Editability · Known limitations · Example prompts · Use cases · Problems solved · Value proposition · Manual alternatives
Connect and manage custom apps for Gemini Apps
View source · Google
Topics: Batch and workflow · Mcp status · Integration ecosystems · Known limitations
Use and manage Connected Apps in Gemini
View source · Google
Topics: Integrations · Integration ecosystems
Take part in the discussion
Sign in to comment, save tools and rate them. You can edit or withdraw your own records. Saved tools are private. Comments are published after moderation and return to review after editing. Rating totals include only valid, approved ratings.
User experiences