To use Gemini, open gemini.google.com or the Gemini mobile app, sign in with a Google Account, type your request in the text box at the bottom and click Submit. The Free plan costs $0 a month with a Google Account and already includes Deep Research, Gemini Live, Canvas, Gems and image generation.
Signed out, the web app runs only some features, and Deep Research and image editing need you to be 18 or older. Paying buys capacity: the free context window is 32k tokens against 1 million on Google AI Pro, and the 5-minute video upload cap rises to 1 hour on Pro or Ultra.
This guide follows a personal Google Account; on a work account, the Workspace edition and admin set the features. Usage refreshes every 5 hours until a weekly limit, and Keep Activity is on by default for adults, so a subset of chats can reach human reviewers.

- Task: set up and use the Gemini app on the web, Android or iPhone
- Result: a signed-in account with files, apps, a research report, an image, a Gem and privacy settings chosen
- Prerequisites: a Google Account; 13 or older for Gems and image generation, 18 or older for Deep Research and image editing
- Tools: gemini.google.com in Chrome, Safari, Firefox, Opera or Edgium, or the Gemini mobile app
- Limitation: usage refreshes every 5 hours until a weekly limit, and the free context window is 32k tokens
In This Article (19 sections)
Before You Begin: What You Need to Use Gemini
You need an eligible Google Account, a supported browser or the mobile app, and the minimum age each feature sets. The Gemini sign-in requirements accept three account types:
- a personal Google Account (accounts managed by Family Link can’t open the web app)
- a work Google Account with a qualifying Workspace edition, if you are 18 or over
- a school Google Account, once the institution’s administrator enables the service
The web app needs Chrome, Safari, Firefox, Opera or Edgium and is available in more than 70 languages; the mobile app runs on some phones. Age and sign-in gate each feature on a personal account:
| Feature | Minimum age | Sign-in required |
|---|---|---|
| Basic chat in the web app | 13 | No, for some features |
| Gems | 13 | Yes |
| Image generation | 13 | Yes |
| Gemini Live | Over 13 | Yes |
| Image editing | 18 | Yes |
| Deep Research | 18 | Yes |
| Customize (connectors and skills) | 18 | Yes |
| Google AI Plus, Pro or Ultra | 18 | Yes |
Sources: image help, Customize help and the step pages linked below, checked Nov 2026.
Signed in, signed out or Keep Activity off: what each mode allows
The mode decides what Google keeps and which features work:
| Mode | Saved to Activity | How long Google keeps chats | Connected Apps | Used to train Google’s AI |
|---|---|---|---|---|
| Signed out (web app) | No | Google Privacy Policy applies; not tied to your account | Not available | Google Privacy Policy applies |
| Signed in, Keep Activity on | Yes | 18 months by default; reviewed chats up to 3 years | Available | Yes |
| Signed in, Keep Activity off | No | Up to 72 hours | Only Device assistance, Gemini Notebook, Messages, Phone and WhatsApp | No, unless you send feedback |
Sources: Gemini Apps Privacy Hub, activity settings and Connected Apps help, checked Nov 2026.
For confidential work, Keep Activity off or a Temporary Chat is the safer default: both keep chats out of model training. The price: with Keep Activity off, Gmail and Drive disappear from Gemini on the web, and past Deep Research reports can’t be found.
Step 1: Open Gemini on the Web, Android or iPhone and Sign In
Open gemini.google.com and sign in from the top right corner:
- On your computer, go to gemini.google.com.
- At the top right, select Sign in.
- Sign in to your Google Account.
On an Android phone, iPhone or iPad, open the Gemini mobile app and sign in with the same account; Gemini Live (Step 9) runs only there.
Expected result: your profile picture or initial appears at the left of the web app, and the text box at the bottom is ready. An error message instead is covered in Troubleshooting.
Step 2: Send Your First Prompt
A prompt takes three actions: type it, optionally attach a file, and send it.
- In the text box at the bottom, enter your question or prompt.
- Optionally, click Add files to attach a document or image.
- Click Submit.
Expected result: Gemini answers, and with Keep Activity on, the chat appears under Recent in the side panel.
Gemini may suggest inaccurate information, so check names, numbers and dates in an answer before you send it on. To keep an output outside Gemini, Deep Research reports (Step 6) and generated images (Step 7) go to Google Docs with Export to Docs.
Hover over a chat under Recent and click More to Pin, Rename or Delete it. A long chat uses more of your limit than a short prompt, because Gemini rereads the history each time, so start a new chat when the topic changes. Once the mechanics work, Gemini prompts for work save you writing each request from scratch.
Use a temporary chat for sensitive questions
A Temporary Chat is for info you don’t want Gemini to remember or use to personalize your experience: Google holds it with your account for 72 hours and does not use it to train its AI models. Use it for a one-off question that names a client.
Step 3: Choose a Model or Leave It on Auto
Leave auto model selection on unless you pay for a Google AI plan: without one, Gemini runs Auto or Flash-Lite, and most auto-routed prompts use Flash-Lite anyway. With a plan, switch models in the web app:
- At the bottom of the chat box, click the current selected model.
- Click the model you want; a lock icon marks models that need an eligible plan upgrade.
- Optionally, select a thinking level: Low, Medium or High (Low is the default).
Expected result: Gemini keeps that model and thinking level for the rest of the conversation, even when you return to it later.
| Plan | Models in chats | Thinking levels |
|---|---|---|
| No Google AI plan | Auto or Flash-Lite | Low, Medium, High |
| Google AI Plus | Flash-Lite, Flash, Pro | Low, Medium, High |
| Google AI Pro | Flash-Lite, Flash, Pro, Argon | Low, Medium, High |
| Google AI Ultra | Flash-Lite, Flash, Argon, Carbon | Low, Medium, High |
Sources: About Gemini models, checked Nov 2026.
Gemini 4 Carbon appears only in the Ultra row. Google calls it its strongest model, designed to reason deeply through your hardest, long-horizon problems, and says Ultra subscribers get it in select countries.
To see which model wrote an answer, click More below the response and select Show thinking steps. Auto selection is switched under Settings > Usage limits.
My take
Stay on auto for everyday prompts. Higher models and thinking levels draw more of your usage, and a paid plan at its limit falls back to Flash-Lite, so I would pick Pro by hand only for complex prompts or coding, Argon only for the complex workflows Google built it for, and Carbon, on Ultra, only for the hardest long-horizon problems.
Step 4: Upload a File and Ask About It
Click Add files, pick the file, and Gemini answers from its content; one prompt takes up to 10 files, subject to availability.
- In the text box at the bottom, enter your question about the file.
- Click Add files, then Upload Files for a file on your device (Add from Drive works once Step 5 connects the Google Workspace app).
- Click Submit.
Expected result: Gemini replies from the file; for an uploaded spreadsheet, you can ask for a chart and change it with Edit.
Drive files need Keep Activity on and Google Workspace connected to Gemini. Under More Uploads, Notebooks adds a Gemini Notebook, which my guide on how to use Gemini Notebook sets up.
The caps that stop an upload:
- 100 MB per file, 2 GB per video
- 5 minutes of total video and 10 minutes of audio, raised to 1 hour and 3 hours on Google AI Pro or Ultra
- one code folder or GitHub repository of up to 5,000 files and 100 MB
A 20-minute meeting recording therefore exceeds the audio cap on Free and on Plus: trim it to 10 minutes or move to Pro.
Note
With Keep Activity on, a subset of chats goes to human reviewers. Remove client names from a file before you upload it, or turn Keep Activity off first (Step 10).
Step 5: Connect Gmail, Drive and Other Apps
Connect an app once under Customize > Connectors, then call it in any prompt by typing @. You must be signed in, 18 or older for Customize, and have Keep Activity on: with it off, Connected Apps are unavailable on gemini.google.com and iOS devices.
- On the sidebar, click Customize (if you don’t find it, first click More).
- At the top, click Connectors.
- Find the app (Gmail and Drive come through the Google Workspace app) and click Add.
- In a new prompt, enter @, select the app, and click Submit.
Expected result: the Connectors tab lists the app as connected, and Gemini uses it when you name it with @. On a work account, the Workspace administrator must enable apps first.
Warning
With your permission, Gemini can edit and manage your content in connected apps. Test the first action on a copy of a file or a draft email, read each action before you confirm it, and use More options > Disconnect for apps you stop using.
Step 6: Run a Deep Research Report
Deep Research turns one question into a sourced report in about 5 to 10 minutes; it needs sign-in and an age of 18 or over.
- In the text box, click Add Files, then Deep Research.
- Optionally, click Sources and add Gmail or Drive; Google Search is included by default.
- Enter what you want researched and click Submit.
- Review the research plan, click Edit plan if it misses something, then click Start research.
- When the report is ready, click Open.
Expected result: the report opens in the Canvas panel, and the web app marks the finished report next to its chat, so you can leave while it runs.
To keep a copy, click Share & export > Export to Docs; without Keep Activity, past reports can’t be found later. Google AI Pro and Ultra get higher report limits and the Pro model, and Ultra reports can add charts and diagrams unless Gmail or Drive is a source.
Step 7: Generate or Edit an Image
Open Images from the sidebar and describe the picture; Nano Banana 2 draws it for any signed-in user 13 or older.
- On the top left, click Open Sidebar, then Images.
- Choose a template or enter a prompt that starts with a word like “create” and names the style you want.
- Hover over the result and click Download full size.
Expected result: the image appears in the chat, and Download full size saves it at 1K without a Google AI plan or 2K with one.
To edit an image (18 and over), open Library, click the image, click Chat under it, describe the change and click Submit. Subscribers can click More > Redo with Pro for a Nano Banana Pro version with more detail in text and infographics.
Google’s Nano Banana page lists Nano Banana 2 as the free model and Nano Banana 2.1, its most advanced image model, as the one a Google AI plan unlocks.
The model setting matters: with Flash-Lite selected, Gemini uses Nano Banana 2 Lite, which isn’t designed for multiple reference images or repeated edits. For trend-style portraits, my Gemini photo prompts are ready to paste.
Step 8: Save a Gem for Repeat Tasks
A Gem saves instructions and reference files so you stop retyping them; starting in November 2026, Gems on personal accounts transition to skills automatically.
- Click Open Sidebar, then Gems (the first time, click Settings and help > Gems).
- Click New Gem.
- Enter a name and instructions.
- Optionally, under “Knowledge,” click Add files to attach reference documents.
- Click Save.
Expected result: the Gem appears under My Gems on the web, in the mobile app and in the Gemini side panel in Google Workspace.
A Drive file added as knowledge stays current, because the Gem reads its most recent version. With the switch this close, keep a copy of your instructions outside Gemini; a skill is created by describing what you want in the prompt bar, and what changes when Gems become skills covers moving existing Gems.
Step 9: Talk to Gemini Live on Your Phone
Gemini Live is a voice conversation in the Gemini mobile app on Android, iPhone or iPad; the web app doesn’t offer it.
- Open the Gemini mobile app (on Android, saying “Hey Google, let’s talk Live” works too).
- At the bottom, tap Live or swipe left.
- Start talking; you can interrupt Gemini mid-answer.
Expected result: Gemini answers out loud, and tapping End (or swiping right) closes Live and shows the transcript.
Hold puts the Live chat on hold and turns off the mic; Mute turns off only your mic while Gemini keeps speaking, so you resume without losing context. On Android, Turn on screen sharing lets Gemini see your screen once notifications for Gemini Live are on. Gems don’t work in Live, and people should agree before you record them in a Live chat.
Step 10: Check Activity and Privacy Settings
Open Settings & help > Activity to see what Gemini saved and decide how long it stays. Keep Activity is on by default if you are 18 or over, and activity older than 18 months is auto-deleted by default.
- Click Settings & help, then Activity.
- To change retention, click Choose an auto-delete option and pick 3, 18 or 36 months, or Don’t auto-delete activity.
- To stop saving chats, near the top click On, then Turn off or Turn off and delete activity.
- To remove past chats, click Delete and pick All time, Last hour, Last day or Custom range.
Expected result: the Activity page shows your Keep Activity setting and auto-delete period near the top.
Chats already reviewed by humans stay for up to 3 years, disconnected from your account, even after you delete activity. On a work or school account, the Workspace administrator controls these settings.
Warning
Deleting a chat also deletes what you created in it, such as Canvas docs, and deleting a Live chat’s activity removes its audio and transcript. Google starts removing deleted activity right away, so back up Deep Research reports with Export to Docs first.
Gemini in Chrome and Google Workspace
Gemini runs inside Chrome and, with the right plan or Workspace edition, inside Gmail and Docs. In Chrome, open a page, click Gemini in Chrome at the top of the browser, enter a prompt, attach anything with Add files > Files, and click Submit.
On a personal account, Google AI Plus adds Gemini in Gmail and Vids, and Pro and Ultra add Docs; Google AI plans are for ages 18 and over.
Work accounts depend on the edition. Business Starter, Business Standard, Business Plus, Enterprise Standard and Enterprise Plus include the Gemini app as a core service, while Business Base and Workspace Individual get it as an additional service the admin must enable. Before asking IT for an upgrade, compare Google Workspace plans and prices.
Free vs Plus vs Pro vs Ultra: Plans, Cost and Limits
Gemini is free: the Free plan costs $0 a month with a Google Account and runs every step above at standard limits, except manual model switching and Redo with Pro. Paid plans raise usage and lift the caps that block long files.

| Plan | US price | Usage limit | Context window | Video / audio upload | Models in chats |
|---|---|---|---|---|---|
| Free | $0 a month | Standard | 32k tokens | 5 min / 10 min | Auto or Flash-Lite |
| Google AI Plus | $4.99 a month | 2x standard | 128k tokens | 5 min / 10 min | Flash-Lite, Flash, Pro |
| Google AI Pro | $19.99 a month or $199.99 a year | 4x standard | 1 million tokens (about 1,500 pages of text) | 1 hour / 3 hours | Flash-Lite, Flash, Pro, Argon |
| Google AI Ultra | $99.99 or $199.99 a month, billed monthly only | 5x or 20x AI Pro | 1 million tokens | 1 hour / 3 hours | Flash-Lite, Flash, Argon, Carbon |
Sources: Google AI plans page, Google One AI plans, Gemini Apps limits and file upload help, checked Nov 2026.
Usage is compute-based: the limit weighs prompt complexity, the models and features you use, and chat length, and it refreshes every 5 hours until you reach a weekly limit.

Settings > Usage Limits shows where you stand. A paid plan at its limit continues on Flash-Lite; Free waits for the refresh and is the first to lose Deep Research when demand peaks. Plus adds video generation and Daily Brief. Annual billing and storage are in my full Gemini pricing breakdown.
My take
Free fits light chat and short files. Plus raises the context to 128k tokens and usage to 2x standard, but not the 5-minute video or 10-minute audio cap. If long recordings are why you would pay, skip Plus and go straight to Pro, the first plan that lifts both the context and the video cap.
How to Verify Gemini Is Set Up Correctly
Each step leaves something visible in the app:
Verification checklist
- Your profile picture or initial shows in the web app instead of a Sign in button
- Your first chat is listed under Recent in the side panel
- Settings & help > Activity lists that chat (Keep Activity on)
- Customize > Connectors shows the Google Workspace app as connected
- Your Deep Research chat holds the finished report with an Open button
- Your Gem appears under My Gems
- Settings > Usage Limits shows your current usage
Practical Example: A First Week With Gemini for a Small Team
Take an illustrative three-person agency on personal Google Accounts, all over 18, starting on the Free plan:
| Task | Feature | Plan needed | Limit to watch | Expected result |
|---|---|---|---|---|
| Summarize a 30-page vendor PDF | Add files > Upload Files | Free | 100 MB per file; 32k-token context | A summary of the terms |
| Summarize client emails | @ Gmail through the Google Workspace app | Keep Activity on | No Connected Apps on the web with Keep Activity off; with it on, a subset of chats can reach human reviewers | A summary of the emails from Gmail |
| Scan three competitors | Deep Research | Free, 18+ | May be unavailable at peak demand | Report exported to Docs |
| Keep proposals on-brand | A Gem with a style guide under Knowledge | Free, 13+ | Moves to skills from November 2026 | Gem under My Gems |
| Review a 20-minute client call | Audio upload | Google AI Pro | Free allows 10 minutes of audio | Call summary |
Sources: Deep Research help and Gems help, checked Nov 2026.
Only the 20-minute call forces a paid plan, because it runs past the 10-minute free audio cap; the rest stays on Free while Keep Activity is on.
Common Mistakes
These defaults and caps are easy to miss:
- Turning Keep Activity off, then losing features. Connected Apps vanish on the web and past reports can’t be found; use a Temporary Chat for the sensitive question instead.
- Uploading a long video on Free. Video caps at 5 minutes; trim it or move to Pro for 1 hour.
- Pasting client data with Keep Activity on. Reviewed chats are kept for up to 3 years, so strip names first.
- Looking for Live on a laptop. It runs only in the mobile app.
- Building Gems without a backup. Personal-account Gems become skills from November 2026.
- Deleting activity before exporting. Canvas docs and Live recordings go with it.
Troubleshooting
Google’s own error messages point to the cause in most cases:
| Problem | Possible cause | What to check |
|---|---|---|
| “Can’t access this service” | Age (13+ personal, 18+ work), a Family Link account, or no admin access | Use a personal account you manage, or ask your admin |
| “Something went wrong” at sign-in | Location, age or account type | Try again later; for age or account type, check the account list in Before You Begin |
| “Delete data to upload file” | Gemini Apps storage for saved activity is full | Delete older activity under Settings & help > Activity |
| “You’ve reached your limit for chats with files” | Rolling upload limit | Wait for the reset, or upgrade to Pro or Ultra |
| “Your uploads may be too large for the best results” | The file is too large for the context window | Upload smaller files, or use Pro or Ultra |
| Usage limit reached | The 5-hour or weekly limit | Settings > Usage Limits shows your usage; a notification gives the refresh time |
| Gmail or Drive missing | Keep Activity off, or apps not enabled by a work admin | Turn Keep Activity on, or contact the admin |
| Recent chats missing | Keep Activity off on a personal account | Turn Keep Activity on |
Sources: file upload help and recent chats help, checked Nov 2026.
FAQ
Can I use Gemini without a Google Account?
Yes, for some features in the web app. Saving activity, file uploads, images, Gems, Connected Apps, Deep Research and Gemini Live all need a signed-in account.
What are the disadvantages of using Gemini?
Gemini can make mistakes, including about people, so double-check facts before you rely on them. Free caps context at 32k tokens, usage runs to a weekly cap, and with Keep Activity on, some chats reach human reviewers. The free plan is still worth it when your use is light chat; for client files, settle Keep Activity before the first upload. My Gemini review weighs whether paying fixes enough.
Does Gemini work in other languages?
Yes. The web app supports more than 70 languages, and the Gemini app is available in 45+ languages; the mobile app doesn’t currently support Farsi.
Can I delete a single chat?
Yes. In the side panel, hover over the chat, click More and then Delete; this removes it from Gemini Apps Activity too. Work and school accounts can’t delete chats.
Next Steps After You Get Started
Setup is finished once every checklist item shows; the next page depends on your job.
Gemini prompts for work
Copy-ready prompts for anyone who has finished setup and wants better answers without writing every request from scratch.
- If Gemini’s limits or privacy defaults don’t fit your work, compare the Gemini alternatives.
- To weigh other assistants by job and budget, start with my ranking of the best AI chatbots.




