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. Nodes
    5. AI Nodes

    AI Nodes

    Integrate artificial intelligence into your workflows

    Overview

    AI nodes bring the power of artificial intelligence to your automations. Use them to generate content, analyze data, make decisions, and even control the browser with human-like vision and reasoning.

    AI Credits
    AI nodes consume AI credits based on usage. Each API call uses credits proportional to the model and tokens used. Monitor your credit usage in Settings → Usage.

    Featured: Prompteryx Browser Agent

    Prompteryx Browser Agent
    When traditional selectors fail, let AI take over. The Prompteryx Browser Agent uses computer vision to see and interact with web pages like a human would.

    When to Use

    • Selectors are unreliable or constantly changing
    • Complex pages with dynamic content
    • Tasks that require reasoning (e.g., "find the cheapest option")
    • Pages with iframes, shadow DOM, or canvas elements
    • When you need human-like judgment in the automation

    Key Features

    • Uses AI vision to find elements - no selectors needed
    • Can make decisions based on page content
    • Works on any page, including complex web apps
    • Choice of model — the Prompteryx models, plus Claude and GPT on cloud runs
    • Runs from the AI Browser Agent page can be turned into Visual Studio nodes (Refine in Visual Studio) for replay without AI costs
    View full documentation →

    All AI Nodes

    Prompteryx Browser Agent
    Prompteryx Browser Agent
    AI-powered browser automation that sees and interacts with pages like a human. Uses computer vision to find elements when selectors fail.

    Common Use Cases

    Complex pages, dynamic content, when selectors are unreliable, tasks requiring reasoning

    View documentation →
    Agent Decision
    Ask a yes/no question about an earlier result and get a reliable true/false (plus answer, reasoning and confidence) to feed into an If Condition. The optional "Jev — instant decision (beta)" model is a decision model rather than a chat model: it answers in well under a second and also returns a calibrated probability (0–1) that the answer is yes. Jev cannot explain its reasoning or extract a value, so its reasoning output only states that probability; if Jev is unavailable the node falls back to the default AI model automatically.

    Common Use Cases

    Branching on an agent's answer, stock / price checks, triage (urgent or not), validating scraped results

    OpenAI / ChatGPT
    OpenAI / ChatGPT
    Chat with GPT-6 Astra / Sol / Luna and GPT-5.x, create images with GPT Image 2.5, speak with GPT-4o mini TTS (13 voices, steerable instructions) and transcribe with GPT Transcribe or Whisper.

    Common Use Cases

    Text generation, summarization, classification, image generation

    View documentation →
    Anthropic / Claude
    Anthropic / Claude
    Claude Fable 5.1, Opus 5.5, Sonnet 5.5, Haiku 5.5, Haiku 4.5 and more. Nodes: Message Claude (system prompt, effort, thinking, JSON output), Analyze image (Claude), Analyze document (Claude) for PDFs, Summarize text, Classify text, Extract data as JSON, Count tokens, and Anthropic file upload/list/get/delete.

    Common Use Cases

    Replies and drafting, summarizing emails, triaging tickets into categories, pulling fields out of invoices, reading PDFs and screenshots

    Google Gemini
    Google Gemini
    Gemini 3.8 Flash, 3.1 Pro and more. Nodes: Generate text (system prompt, JSON output with a schema), Extract data as JSON, Analyze an image / audio / document / video (URL, YouTube link or a file from a previous node), Transcribe a recording (Gemini 3.5 Transcribe), Generate / Edit an image (Nano Banana 2, Pro), Generate a video (Veo 3.1) + Download a video, Text to speech (Gemini TTS, 30 voices), Create embeddings, Count tokens, and Gemini file upload/list/get/delete.

    Common Use Cases

    Reading PDFs and screenshots, transcribing calls, pulling fields out of emails, making social images and voice-overs, short product videos, semantic search

    Perplexity AI
    Perplexity AI
    Search-augmented AI that provides answers with real-time web information. Available through Connect Hub — connect your Perplexity account.

    Common Use Cases

    Research, fact-checking, current information queries

    View documentation →
    ElevenLabs
    ElevenLabs
    Generate natural-sounding speech from text using AI voice synthesis.

    Common Use Cases

    Text-to-speech, voiceovers, audio content generation

    Combining AI with Browser Automation

    AI nodes work seamlessly with browser automation. Common patterns include:

    Scrape → Analyze → Act

    Extract text from a page, use OpenAI to analyze or classify it, then take action based on the result.

    AI Agent Fallback

    Start with selector-based automation, but use Prompteryx Browser Agent as a fallback when selectors fail.

    Generate Content → Fill Forms

    Use AI to generate personalized content, then fill it into web forms automatically.

    Related Documentation

    AI Browser Agent

    Full AI-powered browsing

    Browser Nodes

    Traditional automation

    Variables Guide

    Use AI outputs in workflows

    If Condition

    Branch based on AI responses