Integration

Hyperbrowser integration

Hyperbrowser runs cloud browsers for AI agents. Fetch a page as markdown, HTML, links, a screenshot, JSON extracted by AI or a brand profile, search the web, crawl a site, scrape one or many URLs, extract structured data from pages with a prompt or JSON schema, and run browser agents (HyperAgent, Browser Use, Claude, OpenAI CUA, Gemini, Meta, Grok and Jev computer use) on a natural-language task; list, inspect and stop browser sessions, and manage saved browser profiles. Every call spends credits (1,000 credits = $1; fetch and scrape 1 credit per page, search 5 per query, browser time 100 per hour, proxy data 10,000 per GB, agent steps or model tokens on top). Not exposed: creating raw sessions and their live, CDP, recording and download URLs, manual CAPTCHA evaluation, WebMCP page tools, secrets, vaults and credential fill, sandboxes, volumes and images, extension upload, and X402 wallet payments. Concurrent sessions depend on the plan (Free 1, Startup 25, Scale 100); proxies and CAPTCHA solving need a paid plan. Authentication is a Hyperbrowser API key sent in the x-api-key header.

Connect Hyperbrowser How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Hyperbrowser.Built from the 57 Hyperbrowser tools below, on the connection you add once.

What you can do with Hyperbrowser

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Status, 13 operations

get_batch_scrape_statusGet batch scrape statusGETReturns only the status of a batch scrape job (pending, running, completed, failed), without its pages.
get_browser_use_task_statusGet browser use task statusGETReturns only the status of a Browser Use task (pending, running, completed, failed or stopped), without its result.
get_claude_computer_use_task_statusGet claude computer use task statusGETReturns only the status of a Claude computer use task (pending, running, completed, failed or stopped), without its result.
get_crawl_statusGet crawl statusGETReturns only the status of a crawl job (pending, running, completed, failed), without its pages.
get_extract_job_statusGet extract job statusGETReturns only the status of an extract job (pending, running, completed, failed), without its data.
get_gemini_computer_use_task_statusGet gemini computer use task statusGETReturns only the status of a Gemini computer use task (pending, running, completed, failed or stopped), without its result.
get_grok_computer_use_task_statusGet grok computer use task statusGETReturns only the status of a Grok computer use task (pending, running, completed, failed or stopped), without its result.
get_hyperagent_task_statusGet hyperagent task statusGETReturns only the status of a HyperAgent task (pending, running, completed, failed or stopped), without its result.
get_jev_computer_use_task_statusGet jev computer use task statusGETReturns only the status of a Jev computer use task (pending, running, completed, failed or stopped), without its result.
get_meta_computer_use_task_statusGet meta computer use task statusGETReturns only the status of a Meta computer use task (pending, running, completed, failed or stopped), without its result.
get_openai_cua_task_statusGet openai cua task statusGETReturns only the status of a OpenAI CUA task (pending, running, completed, failed or stopped), without its result.
get_scrape_job_statusGet scrape job statusGETReturns only the status of a scrape job (pending, running, completed, failed), without its data.
get_web_crawl_statusGet web crawl statusGETReturns only the status of a web crawl job (pending, running, completed, failed or stopped), without its results.

Stop, 9 operations

stop_browser_use_taskStop browser use taskPUTStops a running Browser Use task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_claude_computer_use_taskStop claude computer use taskPUTStops a running Claude computer use task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_gemini_computer_use_taskStop gemini computer use taskPUTStops a running Gemini computer use task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_grok_computer_use_taskStop grok computer use taskPUTStops a running Grok computer use task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_hyperagent_taskStop hyperagent taskPUTStops a running HyperAgent task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_jev_computer_use_taskStop jev computer use taskPUTStops a running Jev computer use task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_meta_computer_use_taskStop meta computer use taskPUTStops a running Meta computer use task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_openai_cua_taskStop openai cua taskPUTStops a running OpenAI CUA task for real; it ends with status stopped and its browser stops billing unless keepBrowserOpen was set.
stop_sessionStop sessionPUTStops a browser session for real, ending its billing; the browser closes and unsaved profile changes are lost.

Crawl, 4 operations

crawl_siteCrawl sitePOSTCrawls a site (scraping API): 1 credit per page, 10 with a proxy. Returns the pages with metadata and outputs; read later batches with get_crawl. Waits for the job and returns the finished result; gives up after 10 minutes.
get_crawlGet crawlGETReturns a crawl job by id (from crawl_site): status, totalCrawledPages, totalPageBatches, currentPageBatch and one batch of crawled pages. Use page to read further batches.
get_web_crawlGet web crawlGETReturns one batch of a web crawl's results by job id (from start_web_crawl): status, totalPages, totalPageBatches, currentPageBatch and the pages with their outputs. Use page to read further batches.
start_web_crawlStart web crawlPOSTCrawls a site from a URL, following links: costs per page as fetch_page. Returns each page's outputs (fetch_page formats) and batch counts. Later batches: get_web_crawl. Waits for the job and returns the finished result; gives up after 10 minutes.

Profile, 4 operations

create_profileCreate profilePOSTCreates a saved browser profile (cookies, storage and other browser state) with an optional name. Use its id in sessionOptions.profile with persistChanges true to save a session's state into it.
delete_profileDelete profileDELETEDeletes a saved browser profile for real; the browser state stored in it is lost.
get_profileGet profileGETReturns one browser profile by id (from list_profiles or create_profile): name, team and timestamps.
list_profilesList profilesGETLists saved browser profiles with id, name and timestamps. Filter by name; page with page and limit.

Batch, 2 operations

batch_scrapeBatch scrapePOSTScrapes a list of URLs: 1 credit per page, 10 with a proxy. Returns each page's outputs in the order sent; read later batches with get_batch_scrape. Waits for the job and returns the finished result; gives up after 10 minutes.
get_batch_scrapeGet batch scrapeGETReturns a batch scrape job by id (from batch_scrape): status, totalScrapedPages, totalPageBatches, currentPageBatch and one batch of scraped pages. Use page to read further batches.

Browser use, 2 operations

get_browser_use_taskGet browser use taskGETReturns a Browser Use task by job id (from run_browser_use_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_browser_use_taskRun browser use taskPOSTRuns a Browser Use browser agent on a task and returns finalResult. Acts on websites for real. Costs 20 credits per step plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Claude computer use, 2 operations

get_claude_computer_use_taskGet claude computer use taskGETReturns a Claude computer use task by job id (from run_claude_computer_use_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_claude_computer_use_taskRun claude computer use taskPOSTRuns a Claude computer use browser agent on a task and returns finalResult. Acts on websites for real. Costs model tokens plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Cua, 2 operations

get_openai_cua_taskGet openai cua taskGETReturns a OpenAI CUA task by job id (from run_openai_cua_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_openai_cua_taskRun openai cua taskPOSTRuns a OpenAI CUA browser agent on a task and returns finalResult. Acts on websites for real. Costs model tokens plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Extract, 2 operations

extract_dataExtract dataPOSTExtracts structured data from pages with AI, following a prompt or JSON schema: costs 0.03 credits per output token plus page loads. Returns jobId, status and data. Waits for the job and returns the finished result; gives up after 10 minutes.
get_extract_jobGet extract jobGETReturns an extract job by id (from extract_data): status (pending, running, completed, failed), the extracted data and any error.

Gemini computer use, 2 operations

get_gemini_computer_use_taskGet gemini computer use taskGETReturns a Gemini computer use task by job id (from run_gemini_computer_use_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_gemini_computer_use_taskRun gemini computer use taskPOSTRuns a Gemini computer use browser agent on a task and returns finalResult. Acts on websites for real. Costs model tokens plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Grok computer use, 2 operations

get_grok_computer_use_taskGet grok computer use taskGETReturns a Grok computer use task by job id (from run_grok_computer_use_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_grok_computer_use_taskRun grok computer use taskPOSTRuns a Grok computer use browser agent on a task and returns finalResult. Acts on websites for real. Costs model tokens plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Hyper agent, 2 operations

get_hyperagent_taskGet hyperagent taskGETReturns a HyperAgent task by job id (from run_hyperagent_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_hyperagent_taskRun hyperagent taskPOSTRuns a HyperAgent browser agent on a task and returns finalResult. Acts on websites for real. Costs 20 credits per step plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Jev, 2 operations

get_jev_computer_use_taskGet jev computer use taskGETReturns a Jev computer use task by job id (from run_jev_computer_use_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_jev_computer_use_taskRun jev computer use taskPOSTRuns a Jev computer use browser agent on a task and returns finalResult. Acts on websites for real. Costs model tokens plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Meta computer use, 2 operations

get_meta_computer_use_taskGet meta computer use taskGETReturns a Meta computer use task by job id (from run_meta_computer_use_task): status (pending, running, completed, failed, stopped), data.finalResult and any error.
run_meta_computer_use_taskRun meta computer use taskPOSTRuns a Meta computer use browser agent on a task and returns finalResult. Acts on websites for real. Costs model tokens plus browser time (100 credits/hour). Waits for the job and returns the finished result; gives up after 10 minutes.

Scrape, 2 operations

get_scrape_jobGet scrape jobGETReturns a scrape job by id (from scrape_page): status (pending, running, completed, failed), the data (metadata, markdown, html, links, screenshot) and any error.
scrape_pageScrape pagePOSTScrapes one page: costs 1 credit (10 with a proxy). Returns markdown by default, or the HTML, links or screenshot asked in scrapeOptions.formats, plus page metadata. Waits for the job and returns the finished result; gives up after 5 minutes.

Session, 2 operations

get_sessionGet sessionGETReturns one browser session by id (from list_sessions): status, start and end times, launch settings (stealth, proxy, CAPTCHA, profile) and creditsUsed.
list_sessionsList sessionsGETLists browser sessions (also those agents and jobs launched) with status, start and end times, launch settings and creditsUsed. Filter by status; page with page.

Fetch, 1 operation

fetch_pageFetch pagePOSTLoads a URL in a cloud browser and returns the outputs asked for in outputs.formats: markdown, HTML, links, a screenshot, AI-extracted JSON or a brand profile. Costs 1 credit per page; JSON +5, ultra stealth or location +9.

List, 1 operation

list_extensionsList extensionsGETLists the Chrome extensions uploaded to the account, with id and name. Pass ids as sessionOptions.extensionIds to load them in a job's browser.

Search, 1 operation

search_webSearch webPOSTSearches the web and returns results with title, url and description. Costs 5 credits per query, 10 with location. Use fetch_page on a result url for the full page.

Use Hyperbrowser from Claude Code, Cursor or Codex

Connect Hyperbrowser in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Hyperbrowser operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Hyperbrowser

Create your workspace
Sign up free. Nothing to install.
Connect Hyperbrowser once
Add your Hyperbrowser connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.