Browser Tool

browser_tool

Composio Browser Tool enables AI Agents and LLMs to automate web interactions, perform web scraping, and conduct automated testing. Use cases include data extraction, form automation, website monitoring, and intelligent web navigation.

AI Agents AI Web Scraping Web del proveedor

Acciones

Triggers

Autenticación

—

OAuth gestionado

Información técnica: el detalle de parámetros, schemas y triggers de esta página está pensado para equipos de integración. Si solo necesitas saber si tu herramienta favorita está disponible, basta con ver la lista de acciones.

Acciones disponibles (18)

Cada acción es una operación que el agente puede ejecutar contra este conector. Haz clic en una acción para ver sus parámetros.

Copy Selected TextBROWSER_TOOL_COPY_SELECTED_TEXTAcción

Copy currently selected text on the page to clipboard - ideal for extracting highlighted content, copying form data, or harvesting visible text selections.

Parámetros de entrada

Sin parámetros.

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Drag and DropBROWSER_TOOL_DRAG_AND_DROPAcción

Execute precise drag and drop operations - essential for file uploads, list reordering, element moving, and complex ui interactions that require drag-based manipulation.

Parámetros de entrada

endXintegerObligatorio
Ending X coordinate for drop operation (where to release the element)
endYintegerObligatorio
Ending Y coordinate for drop operation (where to release the element)
buttonstring
Mouse button for drag operation (almost always 'left' for standard drag-drop, 'right' for special contexts)
startXintegerObligatorio
Starting X coordinate for drag operation (where to grab the element)
startYintegerObligatorio
Starting Y coordinate for drag operation (where to grab the element)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Fetch Webpage ContentBROWSER_TOOL_FETCH_WEBPAGEAcción

Your eyes: get page content for decision-making. use before: actions (find targets, understand state) use to verify: page transitions, major state changes, when actions seem to fail format: html=find elements | markdown=clean content | success rate: 99%

Parámetros de entrada

urlstring
URL to fetch content from. If not provided, uses current page
waitinteger
Time to wait for dynamic content to load (milliseconds)
formatstring
Output format for content (markdown or html)
newPageboolean
Whether to create a new page for the content
idleTtlSecinteger
Optional idle TTL for new sessions (seconds)
returnPartialOnTimeoutboolean
Whether to return partial content if timeout occurs

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Get Clipboard ContentBROWSER_TOOL_GET_CLIPBOARDAcción

Read current content from the system clipboard - essential for data transfer workflows, extracting copied text, and reading user-copied data for processing.

Parámetros de entrada

Sin parámetros.

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Keyboard ShortcutBROWSER_TOOL_KEYBOARD_SHORTCUTAcción

Execute keyboard shortcuts and key combinations - essential for copy/paste, navigation, and application commands that agents need for efficient browser automation.

Parámetros de entrada

keysstring[]Obligatorio
Array of keys to press simultaneously (e.g., ['Ctrl', 'c'] for copy, ['Alt', 'Tab'] for window switching, ['Ctrl', 'Shift', 'I'] for dev tools)
holdTimeinteger
Milliseconds to hold keys down (0=quick press like normal shortcuts, 100+=sustained press for special actions)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Mouse ClickBROWSER_TOOL_MOUSE_CLICKAcción

Precision clicker: manual clicking with coordinates. pattern: fetchwebpage(html) → find element → estimate coordinates → click → verify hints: center buttons ~(640,350) | nav/header ~y=150 | content ~y=300-500 tip: try ±50px if first click misses | success rate: 85%

Parámetros de entrada

xintegerObligatorio
X coordinate to click
yintegerObligatorio
Y coordinate to click
buttonstring
Mouse button to use for clicking

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Mouse Double ClickBROWSER_TOOL_MOUSE_DOUBLE_CLICKAcción

Execute a precise double click at specified screen coordinates - ideal for opening files, selecting text, or activating ui elements that require double click gestures.

Parámetros de entrada

xintegerObligatorio
Horizontal pixel position for double click (0 = left edge, increases rightward)
yintegerObligatorio
Vertical pixel position for double click (0 = top edge, increases downward)
buttonstring
Mouse button for double click action - use 'left' for most UI elements, 'right' for context menus

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Mouse Down (Press and Hold)BROWSER_TOOL_MOUSE_DOWNAcción

Press and hold mouse button at coordinates - use for starting custom drag operations, text selections, or long-press interactions. must be followed by mouseup action to complete.

Parámetros de entrada

xintegerObligatorio
X coordinate to press mouse button down (where to start the press-and-hold operation)
yintegerObligatorio
Y coordinate to press mouse button down (where to start the press-and-hold operation)
buttonstring
Mouse button to press and hold (usually 'left' for selections and drags, 'right' for context menus)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Mouse MoveBROWSER_TOOL_MOUSE_MOVEAcción

Move mouse cursor to precise coordinates without clicking - perfect for triggering hover effects, revealing tooltips, and positioning for subsequent interactions.

Parámetros de entrada

xintegerObligatorio
Target X coordinate for mouse cursor (0 = left edge, increases rightward, typical screen: 0-1440px)
yintegerObligatorio
Target Y coordinate for mouse cursor (0 = top edge, increases downward, typical screen: 0-900px)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Mouse Up (Release Button)BROWSER_TOOL_MOUSE_UPAcción

Release mouse button at coordinates - completes drag operations, text selections, and long-press interactions. should be used after mousedown to finish mouse button sequences.

Parámetros de entrada

xintegerObligatorio
X coordinate to release mouse button (where to end the press-and-hold operation or complete a drag)
yintegerObligatorio
Y coordinate to release mouse button (where to end the press-and-hold operation or complete a drag)
buttonstring
Mouse button to release (should match the button used in previous MouseDown action)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Navigate to URLBROWSER_TOOL_NAVIGATEAcción

Always start here: creates browser session and navigates to url. workflow: navigate() → fetchwebpage() → manual interactions → verify print debugurl to user | success rate: 99%

Parámetros de entrada

urlstringObligatorio
URL to navigate the cloud browser to
idleTtlSecinteger
Optional idle TTL for the session (seconds)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Paste TextBROWSER_TOOL_PASTE_TEXTAcción

Paste text content at the current cursor position - perfect for filling forms, inserting data into text fields, or quick content insertion at focused elements.

Parámetros de entrada

textstringObligatorio
Text content to paste at current cursor location (form data, URLs, structured content)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

AI Perform Web TaskBROWSER_TOOL_PERFORM_WEB_TASKAcción

Ai automation: complex workflows only. when to use: 10+ manual steps | dynamic/unpredictable content when to avoid: simple clicks | forms | navigation | payments strategy: try once → if fails, switch to manual immediately success rate: 40% - always have manual backup ready

Parámetros de entrada

urlstring
Optional URL to perform task on (if not provided, uses current page in existing session)
promptstringObligatorio
Natural language description of the web task to perform autonomously (e.g., 'Log into Gmail and send an email', 'Fill out contact form', 'Find and download the PDF')
aiAgentstring
AI agent engine to use ('browser-use'=robust multi-step automation, 'openai-cua'=OpenAI-optimized tasks)
aiModelstring
Specific AI model to use ('gpt-5'=latest flagship, 'gpt-5-mini'=efficient flagship, 'gpt-5-nano'=ultra-fast, 'gpt-4o'=advanced OpenAI, 'gpt-4o-mini'=efficient, 'gpt-4.1'=enhanced, 'gpt-4.1-mini'=compact enhanced)
aiProviderstring
AI provider for task execution ('openai'=GPT models, 'gemini'=Google AI, 'groq'=fast inference, 'azure'=enterprise, 'xai'=experimental)
idleTtlSecinteger
Optional idle TTL for new sessions if sessionId not provided (seconds)
outputSchemaobject
Optional JSON Schema to structure the output for data extraction tasks (provides structured data instead of text description)
highlightElementsboolean
Show visual highlights during task execution for transparency and debugging (True=show interaction highlights, False=hidden execution)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Screenshot WebpageBROWSER_TOOL_SCREENSHOT_WEBPAGEAcción

Capture high-quality screenshot of any webpage with extensive customization options - perfect for archiving, visual documentation, full-page captures, and cross-device viewport testing.

Parámetros de entrada

urlstringObligatorio
URL of the webpage to capture (can be any publicly accessible website)
waitinteger
Milliseconds to wait after page load for JS/animations (1000=default, 3000+=heavy sites, 5000+=very dynamic)
widthinteger
Browser viewport width in pixels for capture (1280=standard desktop, 1440=large desktop, 375=mobile)
heightinteger
Browser viewport height in pixels for capture (720=standard, 900=tall desktop, 667=mobile)
idleTtlSecinteger
Optional idle TTL for new sessions if sessionId not provided (seconds)
imageQualityinteger
Image quality percentage (100=no compression/max quality, 80=good balance, 50=smaller file size)
scrollAllContentboolean
Scroll through entire page to capture all content (True=captures everything, False=viewport only)
captureFullHeightboolean
Capture entire page height ignoring viewport (True=full page length, False=viewport height only)

Parámetros de salida

datastringObligatorio
Base64 encoded image data
errorstring
Error if any occurred during the execution of the action
mimeTypestringObligatorio
MIME type of the image
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Scroll PageBROWSER_TOOL_SCROLLAcción

Page navigation: smooth scrolling. use: when target element not visible after fetchwebpage() distance: 200px=fine | 400px=sections | 800px=quick traverse always: scroll → fetchwebpage() → verify | success rate: 99%

Parámetros de entrada

xinteger
X coordinate where scroll originates (center of viewport recommended: 640 for 1280px width, 720 for 1440px width)
yinteger
Y coordinate where scroll originates (center of viewport recommended: 360 for 720px height, 400 for 800px height)
stepsinteger
Number of steps to break scroll into for smoothness (1=instant jump, 3-5=smooth animation, 10+=very smooth but slower)
useOsboolean
Use OS-level scrolling (True=more reliable, works with all apps) vs browser scrolling (False=web-specific, may fail on some sites)
deltaXinteger
Horizontal scroll distance in pixels (positive=scroll right, negative=scroll left, usually 0 for vertical scrolling)
deltaYintegerObligatorio
Vertical scroll distance in pixels (positive=scroll down toward bottom, negative=scroll up toward top, typical range: 100-500px per scroll)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Set Clipboard ContentBROWSER_TOOL_SET_CLIPBOARDAcción

Store text content in the system clipboard for later paste operations - perfect for preparing data transfers, staging content for forms, or cross-application data sharing.

Parámetros de entrada

textstringObligatorio
Text content to store in system clipboard (URLs, data, structured text, form content)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Take ScreenshotBROWSER_TOOL_TAKE_SCREENSHOTAcción

Visual verification: capture screenshot of current browser viewport. use: debug ui issues, verify page state, document visual results renders: inline in mcp clients for immediate visual feedback tip: use after page changes to confirm they worked | success rate: 99%

Parámetros de entrada

Sin parámetros.

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not

Type TextBROWSER_TOOL_TYPE_TEXTAcción

Controlled input: human-like typing. pattern: click to focus → typetext() → verify speed: delay=0 (fast) | delay=50 (human-like) | delay=100+ (careful) must focus input field first | success rate: 95%

Parámetros de entrada

textstringObligatorio
Text content to type into the browser (supports unicode, special characters, newlines, form data)
delayinteger
Milliseconds delay between each keystroke for natural human-like typing (0=instant, 50-100=realistic human speed, 200+=slow deliberate typing)

Parámetros de salida

dataobjectObligatorio
Data from the action execution
errorstring
Error if any occurred during the execution of the action
successfulbooleanObligatorio
Whether or not the action execution was successful or not