How to Use Gemini Spark and Task Mode

To use Gemini Spark, you need Google AI Pro or Google AI Ultra on a personal Google Account. You must be 18 or older, have Keep Activity on, be opted in to Personal Intelligence, and live outside the EEA, the UK, Switzerland and Nigeria.

Google originally introduced Gemini Spark at Google I/O on May 19, 2026, as an autonomous background agent running on dedicated cloud virtual machines. In October 2026, Google redesigned the interface, removing the Spark entry from the sidebar and renaming the capability to Task mode.

Users manage their goals by submitting plain instructions, approving security prompts for sensitive actions, and checking execution progress inside the Scheduled dashboard.

How to Use Gemini Spark Step by Step

Google moved the background agent controls directly into the conversational view in October 2026, as reported by 9to5Google. The new Chat and Task selector sits directly above the message box on supported interfaces. Once Task mode is selected, users can enter a project goal and approve its actions.

The following sequence initiates a background task from an active account:

• Confirm an active Google AI Pro or Google AI Ultra subscription on an individual account.
• Verify that Keep Activity is enabled and that you are opted into Personal Intelligence within Google account settings.
• Open the Gemini Mac app or Android app. As of October 13, 2026, the Chat/Task switcher is not yet on iOS or gemini.google.com.
• Click the Task toggle above the prompt box to switch out of standard Chat mode.
• Enter your project objective, specifying target documents, spreadsheets, or connected external services.
• Review and confirm permission prompts before the agent sends communications, modifies files, or accesses external websites.
• Track active runs and review tasks on the Scheduled page.

Account Eligibility and Regional Availability Requirements

As of October 2026, Google restricts Task mode to individual personal accounts on Google AI Pro ($19.99 per month) and Google AI Ultra ($99.99 or $199.99 per month), confirmed on the Google AI plans page. The entry-level Google AI Plus tier ($4.99 per month) and the free Gemini tier do not include background agent access. Work and school accounts managed through Google Workspace are not supported for now.

Regional availability remains restricted across several major markets. According to the Google Help Center, Gemini Spark is not available in the European Economic Area (EEA), the United Kingdom (UK), Switzerland, and Nigeria.

Subscription costs and usage caps for each tier are detailed in the guide to Gemini Spark pricing and the broader analysis of Gemini pricing.

Task Scheduling Rules and Concurrency Limits

Scheduled tasks in Gemini Spark can run at specific calendar times or after an event.

Google imposes strict concurrency boundaries on active background operations. Up to 15 tasks can run at the same time on a single account. If 15 tasks are already processing, a newly triggered schedule will not start.

According to Google's documentation, scheduled runs pause whenever task mode is turned off or when a user device is powered down. Routine usage draws from compute-based quotas that reset every 5 hours, with Google AI Pro providing 4 times the compute volume of the free tier and Google AI Ultra providing up to 20 times the Pro allowance on Google's subscriptions page.

Adding Skills and Connectors

The background agent integrates natively with Google core tools, including Calendar, Docs, Drive, Gmail, Keep, Sheets, Slides, Tasks, Contacts, Photos, Gemini Notebook, and YouTube. It also interacts directly with Google Search, Finance, Flights, Hotels, and Maps to gather external real-time data. Third-party integrations expanded on July 1, 2026, to include Canva, Instacart, and OpenTable, followed by Dropbox and Zillow on July 7, 2026, as documented in the Gemini Apps release notes.

Developers can link custom tools using standard Model Context Protocol (MCP) server web addresses, which Google added on June 29, 2026. On the macOS application, these custom MCP connections are labeled Connectors. Users can also configure Skills, which represent reusable context instructions triggered by typing the "@" symbol into the composer. Skills do not require a paid subscription and function within standard chat sessions.

Press reports say Gem creation and editing was turned off on October 13, 2026, as part of the transition toward Skills.

• Google apps: Direct file and communication operations across Gmail, Docs, Drive, Sheets, Slides, Calendar, Keep, and Tasks.
• Supported third-party platforms: Specialized actions inside Canva, Instacart, OpenTable, Dropbox, and Zillow.
• Custom MCP server endpoints: Custom software integrations connected through standardized Model Context Protocol server URLs.
• Modular Skills library: Custom prompt instructions referenced with the "@" character, with initial setup assistance provided by Spark.

Autonomous Chrome Browsing and Safety Approvals

The Chrome auto browse capability launched first for users located in the United States (US). With user permission, the agent can drive the desktop Chrome browser using saved logins and passwords to complete multi-step research. Google also provides an option to execute web navigation inside an isolated remote cloud browser.

Safety boundaries require manual user confirmation before the agent completes high-impact operations. Gemini Spark stops and prompts the user before sending messages, modifying existing files, purchasing items, filling web forms, or signing in with Google credentials. However, bulk updates to private Google Tasks execute automatically without a confirmation prompt.

Supported Devices and How to Use Gemini Spark Across Operating Systems

Gemini Spark operates across mobile, desktop, and web environments, but feature parity differs between client applications.

Google officially lists the Gemini mobile application, the macOS desktop app, and the web interface at gemini.google.com as supported surfaces. The macOS client gained Task mode on October 13, 2026, as detailed by 9to5Google. Remote control from another device arrived July 14, 2026, per the Gemini Apps release notes. Android mobile users began receiving the updated Task mode toggle through a server-side deployment on October 12, 2026.

Platform availability remains uneven across other operating systems. As of October 13, 2026, the iOS mobile app and the standard web interface do not yet display the new Chat and Task switcher. While Google released a dedicated Windows desktop application on September 10, 2026, covered by 9to5Google, Windows is not currently listed as an officially supported surface for Task mode.

Sample Prompts and Practical Task Automation

The agent generates diverse deliverables, including Portable Document Format (PDF) files, standard text documents, spreadsheets, presentation slides, software code, images, audio, video, and compressed archives.

• Recurring fee audit: "Review my monthly credit card and banking statements stored in Google Drive and Gmail over the last 90 days. Flag any price increases or recurring subscription fees, and compile the findings into a Google Sheets summary."
• Trip planning coordination: "Monitor flight options to Chicago on Google Flights for the first weekend of next month. Cross-reference available hotels on Google Hotels, verify that my Google Calendar has no scheduling conflicts, and generate a draft travel itinerary in Google Docs."
• Meeting action item tracking: "Scan my incoming Gmail messages for project updates from the engineering team. Extract all assigned action items, add corresponding deadlines to Google Tasks, and draft a status briefing document in Google Drive."
• Property search monitoring: "Search active residential listings in Austin on Zillow with three bedrooms under $500,000. Save new matching properties to a Google Sheets tracker and save listing photos to a designated Google Drive folder."
• Research synthesis and document creation: "Read the quarterly industry reports stored as PDF files in my research folder on Google Drive. Summarize key competitive benchmarks into a new Google Docs report and export the executive summary as a Portable Document Format (PDF) file."

Troubleshooting Common Errors and Missing Interface Features

When Task mode fails to appear, check whether you are on iOS or gemini.google.com, where the switcher has not arrived, and whether your account meets the requirements above. Accounts on Google AI Plus or the free tier cannot access the feature, nor can accounts registered to minors or managed by Workspace administrators. Disabling Keep Activity or opting out of Personal Intelligence also blocks Task mode.

Users searching for missing controls should note that Google removed Spark from the left navigation sidebar during the October 2026 interface redesign. Ongoing tasks and recurring schedules now reside on the dedicated Scheduled page. For a full breakdown of the redesign timeline, review the guide explaining what happened to Gemini Spark.

If the interface displays an alert stating "Something went wrong (1076)", note that Google has never published an official definition for error 1076. It represents a general Gemini error rather than a bug unique to Task mode.