PrompteryxBeta
    ⌘K
    PlatformSupport
    Menu
    IntroductionInstall Desktop & extensionQuick StartPlatform Overview
    OverviewWhat you can makeAI Browser Agent in the chatModels and AutoConnectorsSchedulesProjectsExplore and publishingShare linksMemory and SkillsPersonal agent (beta)Welcome, settings and skinsPlans, credits and top-upsPrompteryx DesktopGroups (pilot)Phone app (early access)
    OverviewCreating WorkflowsTesting & DebuggingScheduling
    OverviewGetting StartedCommandsRecordingSettingsStructured OutputParallel AgentsTips & Best Practices
    OverviewInstallationSidebar AgentUsing Attached FilesExecute on Current TabExporting to Visual Studio
    Overview
    OverviewOpen PageClickClick CoordinatesFillExtract TextScreenshotWaitPress KeyFocus ElementBlur ElementSelect TextHighlight ElementSet ViewportGet Bounding BoxScrollHoverSelect OptionAssertionsCookiesEvaluate JavaScriptExample Workflows
    OverviewPrompteryx Browser AgentOpenAI
    OverviewSchedule TriggerWebhook Trigger
    OverviewGoogle SheetsGoogle CalendarConnect Hub (3,000+ apps)Example Workflows
    AI Vision SettingsChrome Automation ProfilesCloud BrowserBrowser Profiles & Persistent ContextVideo RecordingExecution LogsSharing & CollaborationEdit History
    Selectors GuideVariables
    Start hereGet an API keyRun the AI browser agentDrive a cloud browserRun workflows from anywhereUse connected apps (Connect Hub)SDK (TypeScript)CLIMCP for AI assistants
    REST APIWebhooksMCP toolsCloud Browser APICloud Browser quickstartMigrate from Browserbasen8nZapierMake.com

    Need help? Contact Support

    1. Documentation
    3. AI Browser

    AI Browser Agent

    AI-Powered

    An AI browsing copilot with computer vision that runs on our Cloud Browser or in your own Chrome

    Now in the main chat. You can run the AI Browser Agent straight from the app's home: choose AI Browser Agent under the message box, or just ask for something that needs a website. Runs end with an outcome card with Try again and Try with another model. See AI Browser Agent in the chat. The full-page agent described here still works, with the same settings.
    AI Browser Agent

    AI-Powered Browser Automation

    Just prompt it and watch it navigate, click, type, extract, and complete tasks across the web—then convert the run into a reusable workflow.

    The AI Browser Agent workspace: sidebar with chat history, the composer, and the curated Model dropdown open showing the available AI models
    The AI Browser Agent workspace — type a task into the composer, pick a model, and run it on your Chrome or a cloud browser.
    AI Credits Required
    The AI Browser Agent uses AI credits for each action. Credits are consumed based on the complexity of the task and the AI model used. The "AI Credits" pill in the page header shows your balance, and Settings → Usage shows what you've used this month.

    Key Features

    Natural Language Commands
    Delegate multi-step web work with plain English prompts. The agent understands context and handles real sites, forms, and complex flows.
    Computer Vision
    Uses AI vision to "see" the page and interact with elements visually. No need for CSS selectors - the agent finds elements by what they look like.
    Live Transparent Execution
    Watch the agent work in real time in the live view beside the chat. See exactly what it's doing step-by-step as it navigates.
    Convert to Workflows
    Turn the agent's actions into a node-based workflow automatically. Refine each step for reliability, then run without AI costs.
    Video Recording
    Optionally record the entire session for a complete visual trail of every action taken. Great for debugging and auditing.
    Stop or Take Over
    Stop the agent instantly (the Stop button or Esc) whenever you need to intervene. The browser stays open, so a follow-up message corrects direction. You're always in control.

    Quick Navigation

    Overview
    Learn how the AI Browser Agent works and when to use it
    Getting Started
    Start your first AI Browser session and complete a task
    Commands
    Learn effective prompting techniques for the AI agent
    Recording Actions
    Convert AI Browser sessions into reusable Visual Studio workflows

    When to Use AI Browser Agent

    ✓ Best For

    • • One-off complex tasks you don't want to automate
    • • Pages with constantly changing layouts
    • • Tasks that require reasoning (e.g., "find the best deal")
    • • Creating initial workflows to refine later
    • • Handling dynamic content (React, Angular, etc.)
    • • Elements in iframes, shadow DOM, or canvas

    ⚠️ Consider Visual Studio Instead

    • • Repetitive tasks you'll run many times
    • • Simple, predictable page structures
    • • When you want to minimize AI credit usage
    • • Scheduled automations that run unattended
    • • Tasks requiring precise timing or sequences
    Pro Tip: Use the Agent to Build Workflows
    The smartest way to use AI Browser Agent is to let it complete a task once, then save the actions as a Visual Studio workflow. You can then run that workflow repeatedly without consuming AI credits. Think of it as "teaching" the automation once.