Hyperbrowser is a next-generation platform empowering AI agents and enabling effortless, scalable browser automation.
Hyperbrowser is a next-generation platform empowering AI agents and enabling effortless, scalable browser automation. On Nagent, Hyperbrowser is exposed as a fully-configurable ai web scraping integration that any agent can call — 43 actions, and API key authentication. No code is required to wire Hyperbrowser into your workflow — connect it once via the External Integrations panel and reuse it across every agent you build.
Agent builders use Hyperbrowser to automate the kinds of tasks ai web scraping teams previously handled manually. Concrete examples — each one is a single agent step in Nagent — include:
Every action and trigger is paired with a structured input/output schema (visible in the sections below), so when you wire Hyperbrowser into Helix — our agentic agent builder — the editor knows exactly what each step expects and produces. Configure once, deploy anywhere across your Nagent agents.
Every operation an agent can call against Hyperbrowser, with input parameters and output schema. Drop these into any step of an agent built in Helix.
HYPERBROWSER_ADD_EXTENSIONTool to add a new browser extension to Hyperbrowser for use in sessions. Use when you need to upload a Chrome extension zip file that can be loaded into browser sessions. The extension zip must contain a valid manifest.json file at the root level.
Input parameters
Chrome extension zip file to upload. The zip must contain a valid manifest.json at the root level.
Optional display name for the extension. If not provided, the filename will be used as the extension name.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_CREATE_PROFILECreates a new persistent Hyperbrowser profile for storing browser state (cookies, sessions, etc.). Use this to create a reusable profile that can be attached to browser sessions via the profile ID. Profiles allow you to maintain logged-in states across multiple sessions without re-authenticating.
Input parameters
Optional profile name for easy identification. If not provided, an unnamed profile will be created. Must be at least 1 character if specified.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_CREATE_SCRAPE_JOBTool to initiate a new scrape job. Use when you need to extract structured content from a target URL with custom session and scrape settings.
Input parameters
Target URL to scrape
Scrape behavior configuration Key options: `onlyMainContent` (bool, omits boilerplate/nav content), `waitFor` (ms to wait for JS rendering), `timeout` (ms before capture). Set `onlyMainContent=true` for article/focused content; increase `waitFor` and `timeout` for JS-heavy dynamic pages.
Browser session configuration
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_CREATE_SESSIONTool to create a new browser session with custom stealth, proxy, and privacy settings. Use when initializing an automated browsing session with specific configuration.
Input parameters
Preferred device types
Screen configuration: width and height in pixels.
Block ads in the session
Browser locale codes (ISO 639-1 two-letter codes, e.g., 'en', 'fr', 'de', 'zh')
Browser profile reuse settings.
Preferred browser platforms
Block trackers during session
Route session via a proxy server (requires paid plan)
Desired proxy city; mutually exclusive with proxyState
Block common annoyances like pop-ups
Two-letter US state code; mutually exclusive with proxyCity
Enable stealth mode to reduce detection
Additional browser launch arguments
Proxy server host or URL
List of extension IDs (UUIDs) to attach
Two-letter proxy country code (e.g., US, GB)
List of URL patterns to block
Automatically accept cookie banners
Auto-solve CAPTCHAs during session (requires paid plan)
Auto-stop after N minutes (1-720)
Preferred OS fingerprints
Enable rrweb session recording
Image CAPTCHA handling parameters
Password for proxy authentication
Username for proxy authentication
Enable video recording; requires enableWebRecording=true
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_DELETE_PROFILETool to delete a profile. Use when you need to remove a profile by its unique identifier after confirming its existence.
Input parameters
Unique identifier of the profile to delete.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_FETCH_WEB_PAGETool to fetch a web page and return content in various formats (HTML, Markdown, JSON, screenshot, etc.). Use when you need to retrieve and process web content with customizable browser settings, output formats, and content filtering options.
Input parameters
Target URL to fetch
Cache control options
Browser configuration options
Output format and content filtering options
Stealth mode level to avoid detection
Page navigation and loading options
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_BROWSER_USE_TASK_RESULTDEPRECATED: Use HYPERBROWSER_GET_BROWSER_USE_TASK_STATUS instead. Tool to retrieve the status and results of a browser use task. Use after starting a browser use task to check its progress, retrieve execution steps, and obtain the final result.
Input parameters
The jobId of the browser use task. This is the UUID returned when starting a browser use task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_BROWSER_USE_TASK_STATUSTool to retrieve the current status of a browser-use task. Use when checking if a browser automation task has completed or is still pending.
Input parameters
Unique identifier of the browser-use task to query. This is the jobId returned when starting a browser-use task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_CLAUDE_COMPUTER_USE_TASK_RESULTTool to retrieve the complete result and status of a Claude Computer Use task. Use when you need full task details including execution steps, final results, and error information.
Input parameters
The jobId returned when starting a Claude Computer Use task via HYPERBROWSER_START_CLAUDE_COMPUTER_USE_TASK. This is a UUID format string.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_CLAUDE_COMPUTER_USE_TASK_STATUSPoll the execution status of a Claude Computer Use task. Use after calling HYPERBROWSER_START_CLAUDE_COMPUTER_USE_TASK to check if the task has completed, is still running, or has failed. Pass the jobId returned from the start task action.
Input parameters
The jobId returned when starting a Claude Computer Use task via HYPERBROWSER_START_CLAUDE_COMPUTER_USE_TASK. This is a UUID format string.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_CRAWL_JOB_STATUSTool to retrieve the status and results of a specific crawl job. Use after submitting a crawl job to check its progress or fetch results.
Input parameters
Crawl job identifier (UUID)
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_CRAWL_STATUSTool to retrieve the current status of a specific crawl job. Use after initiating a crawl job to poll its status.
Input parameters
UUID of the crawl job to retrieve status for
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_CUA_TASK_RESULTTool to retrieve the status and results of a CUA (Claude User Agent) task. Use after starting a CUA task to check its progress, retrieve execution steps, and obtain the final result.
Input parameters
The jobId of the CUA (Claude User Agent) task. This is the UUID returned when starting a CUA task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_CUA_TASK_STATUSPoll the execution status of a CUA task. Use to check if a CUA task has completed, is still running, or has failed.
Input parameters
The ID of the CUA task to query. This is a UUID format string returned when starting a CUA task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_EXTRACT_JOB_RESULTTool to fetch the status and results of a specific extract job. Use after initiating an extract job to monitor progress and retrieve final data.
Input parameters
The UUID of the extract job returned by the Start Extract Job action
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_EXTRACT_JOB_STATUSRetrieve the status of an extract job. Use after calling Start Extract Job to poll its progress. Poll periodically until status is 'completed' or 'failed', then use Get Extract Job Result to retrieve the extracted data.
Input parameters
UUID of the extract job returned by the Start Extract Job action (e.g., 'd0208f8c-0d1c-4697-bfc8-031f9bec20fc')
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_GEMINI_COMPUTER_USE_TASK_RESULTTool to retrieve the current status and results of a Gemini Computer Use task. Use when checking if a Gemini automation task has completed or is still pending.
Input parameters
Unique identifier of the Gemini Computer Use task to query. This is the jobId returned when starting a Gemini Computer Use task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_HYPER_AGENT_TASK_RESULTTool to retrieve the status and results of a HyperAgent task. Use when checking if a HyperAgent task has completed or to get the final results and execution steps.
Input parameters
The unique identifier (UUID) of the HyperAgent task. This is the jobId returned when starting a HyperAgent task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_PROFILERetrieves details of a specific Hyperbrowser profile by its UUID. Returns profile metadata including name, team association, and timestamps. Use the List Profiles action first to obtain valid profile IDs.
Input parameters
Unique identifier (UUID) of the profile to retrieve. Get profile IDs from the List Profiles action.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_SCRAPE_JOB_RESULTRetrieves the status and results of a scrape job. Poll this endpoint after creating a scrape job to check progress and get the scraped content when completed. Returns jobId, status (pending/running/completed/failed), scraped data, and any error message.
Input parameters
UUID of the scrape job to retrieve (returned by create_scrape_job)
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_SCRAPE_JOB_STATUSTool to retrieve the current status of a specific scrape job. Use after initiating a scrape job to poll its status; poll at moderate intervals (e.g., every 2–5 seconds) rather than tight loops to avoid exhausting quota. Responses may be large when the job includes screenshots or full HTML.
Input parameters
UUID of the scrape job to retrieve status for
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_SESSION_DETAILSRetrieve detailed information about a Hyperbrowser session by its ID. Use this tool to get connection endpoints (wsEndpoint, liveUrl), session status, and configuration details for an existing session. The session ID can be obtained from create_session or list_sessions actions. Returns active connection details (wsEndpoint, liveUrl, token) for active sessions, or historical information for closed sessions.
Input parameters
The unique session ID (UUID format) to retrieve details for. Obtain this from create_session or list_sessions.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_SESSION_DOWNLOADS_URLTool to retrieve the downloads URL for a session. Returns a signed URL to download files saved during a browser session. Note: The session must be created with 'saveDownloads: true' for downloads to be available. Poll this endpoint checking status until 'completed'.
Input parameters
UUID of the session (must be a valid UUID format)
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_SESSION_RECORDINGRetrieve the recording URL for a browser session. Returns a pre-signed S3 URL to download the rrweb JSON recording. The recording status indicates availability: - 'pending': Recording is being prepared - 'in_progress': Recording is still being processed - 'completed': Recording is ready, recordingUrl will contain the download link - 'failed': Recording failed, check the error field - 'not_enabled': Session was created without enableWebRecording=true Poll this endpoint until status is 'completed' or 'failed' after stopping a session.
Input parameters
UUID of the session to retrieve recording for. Must be a valid UUID format (e.g., '550e8400-e29b-41d4-a716-446655440000').
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_SESSION_VIDEO_RECORDING_URLTool to retrieve the video recording URL for a browser session. Returns a pre-signed URL to download the video recording. The video recording status indicates availability: - 'pending': Video recording is being prepared - 'in_progress': Video recording is still being processed - 'completed': Video recording is ready, recordingUrl will contain the download link - 'failed': Video recording failed, check the error field - 'not_enabled': Session was created without enableVideoWebRecording=true Poll this endpoint until status is 'completed' or 'failed' after stopping a session.
Input parameters
UUID of the session to retrieve video recording for. Must be a valid UUID format (e.g., '21de56ee-8ba3-4b07-8e5b-bf0ac94445d1').
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_WEB_CRAWL_RESULTTool to retrieve the status and results of a web crawl job. Use after submitting a web crawl job to check its progress and fetch paginated results. Supports pagination via page and batchSize parameters for large crawl jobs.
Input parameters
Web crawl job identifier (UUID)
Page number for result pagination (1-indexed). Retrieves specific batch of results.
Number of items to return per page batch. Controls pagination size.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_GET_WEB_CRAWL_STATUSTool to retrieve just the status of a web crawl job without the full results. Use after initiating a web crawl to poll its current state.
Input parameters
UUID of the web crawl job to retrieve status for
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_LIST_EXTENSIONSTool to list all browser extensions. Use when you need to fetch all available extensions for the Hyperbrowser account.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_LIST_PROFILESTool to list profiles. Use when you need to fetch paginated profiles and optionally filter by name.
Input parameters
Filter profiles by name containing this substring (case-insensitive)
Page number to retrieve (1-based index)
Number of profiles to return per page
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_LIST_SESSIONSTool to list sessions with optional status filter. Use when you need a paginated overview of browser sessions before acting on them.
Input parameters
Page number for pagination (default 1)
Filter sessions by status
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_SEARCH_WEBTool to perform a web search and retrieve results with titles, URLs, and descriptions. Use when you need to search the web for information on a specific topic or query.
Input parameters
Page number for paginated results (starts at 1)
Search query string
Advanced filters for refining search results.
Geographic location for localized search results.
Maximum age of results in seconds (e.g., 86400 for last 24 hours)
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_BROWSER_USE_TASKTool to start an asynchronous browser-use task. Use when you need to automate web interactions given a task instruction.
Input parameters
Language model to drive the browser automation agent
Natural language description of what the AI agent should accomplish. Be specific and clear about the goal.
Custom API keys for providers
Max agent steps before abort. Maximum allowed value depends on your Hyperbrowser plan (common plans cap at 25 steps)
Reuse an existing browser session ID Useful for CAPTCHA-protected or bot-detection sites where a pre-authenticated or warmed-up session avoids re-triggering challenges.
Enable vision for page context
LLM to use for planning
Max failures allowed before abort
Actions to run before starting the task
Max input tokens allowed
Browser session configuration
Whether to validate the final output
Keep the browser open after completion
Seconds between planner cycles
Use custom provider API keys
Max actions per planning step
LLM to use for page extraction
Allow planner to use vision
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_CLAUDE_COMPUTER_USE_TASKTool to start a Claude Computer Use task. Use when you need AI-driven automated browser interactions. Call after you have your task prompt and any session preferences configured.
Input parameters
Model variant to use for the task
Natural language instructions for the Claude agent
Custom API key configuration.
Max action steps to run
Reuse an existing session by ID
Max allowed failures before abort
Overrides for browsing session settings.
Keep browser open after completion
Whether to supply custom API keys
Enable built-in Claude computer action library
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_CRAWL_JOBTool to start a new crawl job for a specified URL. Use when you need to initiate a web crawl before checking job status.
Input parameters
Starting URL for the crawl job
Maximum pages to crawl (default: 10, max: 100)
Whether to follow links during crawl
Ignore sitemap.xml files if present
Configuration for post-crawl data extraction.
Settings to configure the browser session during crawl.
URL patterns to exclude from crawling
URL patterns to include in crawling
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_CUA_TASKTool to start an OpenAI CUA (Computer-Using Agent) task. Use when you need AI-driven automated browser interactions powered by OpenAI. Call after you have your task prompt and any session preferences configured.
Input parameters
Natural language instructions for the OpenAI CUA agent
Custom API key configuration.
Max action steps to run
Reuse an existing session by ID
Max allowed failures before abort
Overrides for browsing session settings.
Keep browser open after completion
Whether to supply custom API keys
Enable built-in computer action library
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_EXTRACT_JOBStart an AI-powered data extraction job from one or more web pages. Use this tool to scrape structured data from websites by providing URLs and either a natural language prompt describing what to extract, or a JSON schema defining the output structure (or both for best results). Returns a jobId to track extraction progress via HYPERBROWSER_GET_EXTRACT_JOB_STATUS and retrieve results via HYPERBROWSER_GET_EXTRACT_JOB_RESULT.
Input parameters
List of URLs to extract from (maximum 10 URLs per request); append /* to enable crawling the entire site
Natural language instructions describing what data to extract from the page(s). Either prompt or schema must be provided.
JSON schema defining the structure of extracted data. Either prompt or schema must be provided. For best results, provide both.
Milliseconds to wait after page load before extraction
Max links to consider when crawling (for URLs ending with /*)
Optional system prompt to guide the extractor
Browser/session configuration options
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_GEMINI_COMPUTER_USE_TASKTool to start a Gemini Computer Use task for browser automation using Google's Gemini. Use when you need AI-driven automated browser interactions with Gemini models.
Input parameters
Model variant to use for the task
Natural language instructions for the Gemini agent
Custom API key configuration.
Max action steps to run
Reuse an existing session by ID
Max allowed failures before abort
Overrides for browsing session settings.
Keep browser open after completion
Whether to supply custom API keys
Enable built-in Gemini computer action library
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_START_WEB_CRAWLTool to start an asynchronous web crawl job that follows links from a starting URL and returns content from each page. Use when you need to crawl multiple pages from a website.
Input parameters
Starting URL for the web crawl
Caching configuration for crawl results.
Browser configuration options.
Configuration for content extraction and output.
Stealth mode options for browser fingerprinting.
Navigation timing and wait configuration.
Options for controlling the web crawl behavior.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_STOP_BROWSER_USE_TASKTool to stop a running browser-use task. Use when halting an in-progress browser automation task after confirming its task ID.
Input parameters
Unique identifier of the browser-use task to stop
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_STOP_CLAUDE_COMPUTER_USE_TASKTool to stop a running Claude computer use task. Use when a Claude computer use task is in progress and needs to be terminated.
Input parameters
Unique identifier (jobId) of the Claude computer use task to stop. This ID is returned when starting a task via the Start Claude Computer Use Task action.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_STOP_CUA_TASKTool to stop a running CUA task. Use when a CUA task is in progress and needs to be terminated.
Input parameters
Unique identifier of the CUA task to stop. This ID is returned when starting a CUA task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_STOP_GEMINI_COMPUTER_USE_TASKTool to stop a running Gemini computer use task. Use when a Gemini computer use task is in progress and needs to be terminated.
Input parameters
Unique identifier (jobId) of the Gemini computer use task to stop. This ID is returned when starting a task via the Start Gemini Computer Use Task action.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
HYPERBROWSER_STOP_SESSIONTool to stop a running session by ID. Use after confirming the session is active.
Input parameters
Unique identifier of the session to stop
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
No publicly available marketplace agent is found using this tool yet. There are 34 agents privately built on Nagent that already use Hyperbrowser.
Build on Nagent
Connect Hyperbrowser to any Nagent agent in minutes — no API key management, no boilerplate. Just configure and deploy.
The five questions agent builders ask before adopting a new integration.
Open the External Integrations panel inside Nagent (app.nagent.ai/externalIntegration), find Hyperbrowser, and click "Connect Now." You'll authenticate with an API key — Nagent handles credential storage and refresh automatically. Once connected, Hyperbrowser is available to any agent in your workspace.
No. Nagent provides no-code integration for every tool. Once Hyperbrowser is connected, you configure its 43 actions directly in the agent builder UI — no API calls, no boilerplate, no schema management.
Helix — Nagent's agentic agent builder — lets you drop Hyperbrowser steps into any workflow visually. Pick an action (e.g., one of those listed above), fill in the inputs (Helix knows the required vs. optional schema for each parameter), and connect it to upstream/downstream steps. Triggers run as the entry point of an agent, so when a Hyperbrowser event fires, the agent kicks off automatically.
Every Hyperbrowser action and trigger ships with a fully-typed schema — input parameters with name, type, required flag, and description, plus the output payload shape. The schemas are documented in the sections above. Helix uses these schemas to validate your configuration at build time and to type-check the data flowing between steps.
Yes. While Hyperbrowser ships with 43 pre-built ai web scraping actions, you can layer custom logic around them inside Helix — pre/post-processing steps, conditional branches, retries, or stitching Hyperbrowser together with other connected tools. For deeper customization, talk to our team about Nagent's Agentic AI Lab — forward-deployed engineers who build Hyperbrowser-based workflows tailored to your business.