> **Building with AI coding agents?** Install the authstack plugin with one command. This equips your agent with accurate Scalekit implementation patterns.
>
> **Recommended** (the `-y` flags skip prompts, so the command runs without an interactive terminal):
> ```bash
> npx -y @scalekit-inc/cli setup -y
> ```
>
> Global:
> ```bash
> npm install -g @scalekit-inc/cli
> scalekit setup -y
> ```
>
> Supports Claude Code, Cursor, GitHub Copilot, Codex + skills for other Agent Skills-compatible agents in `~/.agents/skills`.
> Skills: integrate-agentkit, implement-saaskit, add-mcp-oauth, implement-sso, implement-scim.
> [Full setup guide](https://docs.scalekit.com/dev-kit/build-with-ai/)

---

# Connect AI agents to Hyperbrowser

The Hyperbrowser connector lets your AI agent act in each user's Hyperbrowser account. Each user connects their own Hyperbrowser API key once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 60 tools.

**Authentication:** API Key
**Categories:** Automation, Developer Tools
**Tools:** 60: 35 read, 16 write, 9 destructive
**Users sign in with:** API key
**Built by:** Scalekit connector
**Try it:** [Playground in the Scalekit dashboard](https://app.scalekit.com/ws/signup?sk_intent=playground&provider=HYPERBROWSER)

## Setup

1. ### Install the SDK

   **Node.js**

   ```bash
   npm install @scalekit-sdk/node dotenv
   ```

   **Python**

   ```bash
   pip install scalekit-sdk-python python-dotenv
   ```

2. ### Set your credentials

   Add your Scalekit credentials to your `.env` file. Find values in **[app.scalekit.com](https://app.scalekit.com)** > **Developers** > **API Credentials**.

   ```sh title=".env"
   SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
   SCALEKIT_CLIENT_ID=<your-client-id>
   SCALEKIT_CLIENT_SECRET=<your-client-secret>
   ```

3. ### Create the Hyperbrowser connection

   In **AgentKit > Connections**, create a Hyperbrowser connection. The name you give it is the `connection_name` your code passes. See [Configure connections](/agentkit/connections/).

4. ### Authorize a user and make your first call

   **Node.js** (`quickstart.mts`)

   ```typescript
   import { ScalekitClient } from '@scalekit-sdk/node'
   import 'dotenv/config'
   import { createInterface } from 'node:readline/promises'

   const scalekit = new ScalekitClient(
     process.env.SCALEKIT_ENVIRONMENT_URL,
     process.env.SCALEKIT_CLIENT_ID,
     process.env.SCALEKIT_CLIENT_SECRET,
   )
   const actions = scalekit.actions

   const connector = 'hyperbrowser'
   const identifier = 'user_123'

   // Generate an authorization link for the user
   const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
   console.log('Authorize Hyperbrowser:', link)
   const rl = createInterface({ input: process.stdin, output: process.stdout })
   await rl.question('Press Enter after authorizing...')
   rl.close()

   // Make your first call
   const result = await actions.executeTool({
     connector,
     identifier,
     toolName: 'hyperbrowser_extensions_list',
     toolInput: {},
   })
   console.log(result)
   ```

   ```bash
   npx tsx quickstart.mts
   ```

   **Python** (`quickstart.py`)

   ```python
   import os
   from scalekit import ScalekitClient
   from dotenv import load_dotenv
   load_dotenv()

   scalekit_client = ScalekitClient(
       env_url=os.getenv("SCALEKIT_ENVIRONMENT_URL"),
       client_id=os.getenv("SCALEKIT_CLIENT_ID"),
       client_secret=os.getenv("SCALEKIT_CLIENT_SECRET"),
   )
   actions = scalekit_client.actions

   connection_name = "hyperbrowser"
   identifier = "user_123"

   # Generate an authorization link for the user
   link_response = actions.get_authorization_link(
       connection_name=connection_name,
       identifier=identifier,
   )
   print("Authorize Hyperbrowser:", link_response.link)
   input("Press Enter after authorizing...")

   # Make your first call
   result = actions.execute_tool(
       tool_input={},
       tool_name="hyperbrowser_extensions_list",
       connection_name=connection_name,
       identifier=identifier,
   )
   print(result)
   ```

   ```bash
   python quickstart.py
   ```

   Each user opens the link once and enters their Hyperbrowser credentials there. See [Authorize a user](/agentkit/tools/authorize/) for the full flow and statuses.

## Tools

Pass the exact name to `execute_tool`, with an input like each tool's example. To ask for a tool that's missing, use the [request form](https://scalekitsupport.portal.usepylon.com/forms/request-a-connector-tool).

**Node.js**

```typescript
const result = await actions.executeTool({
  toolName: 'hyperbrowser_browser_use_task_get',
  toolInput: {
    task_id: '<task_id>',
  },
  connector: 'hyperbrowser',
  identifier: 'user_123',
})
```

**Python**

```python
result = actions.execute_tool(
    tool_name="hyperbrowser_browser_use_task_get",
    tool_input={
        "task_id": "<task_id>",
    },
    connection_name="hyperbrowser",
    identifier="user_123",
)
```

`result.data` is the app's response as JSON, and `result.execution_id` (`executionId` in Node.js) is the ID of the call.

### `hyperbrowser_browser_use_task_get`

Get Browser Use Task · Read-only

Get a Browser Use task's status, steps and final result. Returns the task object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_browser_use_task_status_get`

Get Browser Use Task Status · Read-only

Get only the status of a Browser Use task. Returns the status. Use hyperbrowser_browser_use_task_get for the result.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_claude_computer_use_task_get`

Get Claude Computer Use Task · Read-only

Get a Claude Computer Use task's status, steps and final result. Returns the task object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_claude_computer_use_task_status_get`

Get Claude Computer Use Task Status · Read-only

Get only the status of a Claude Computer Use task. Returns the status. Use hyperbrowser_claude_computer_use_task_get for the result.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_crawl_get`

Get Crawl Job · Read-only

Get a crawl job's status and crawled pages, a page of results at a time. Returns the job with page results. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `crawl_job_id` (`string`, required): ID of the crawl job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `batch_size` (`integer`, optional): Number of page results per response page. Example: 20.
- `page` (`integer`, optional): Page number of results, starting at 1. Example: 1.

Example input: `{"crawl_job_id":"<crawl_job_id>"}`

### `hyperbrowser_crawl_status_get`

Get Crawl Job Status · Read-only

Get only the status of a crawl job. Returns the status. Use hyperbrowser_crawl_get for the pages.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `crawl_job_id` (`string`, required): ID of the crawl job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"crawl_job_id":"<crawl_job_id>"}`

### `hyperbrowser_extensions_list`

List Extensions · Read-only

List the Chrome extensions uploaded to the account. Returns extension IDs and names. Pass IDs as extension_ids to hyperbrowser_session_create; extensions are uploaded in the dashboard.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `hyperbrowser_extract_get`

Get Extract Job · Read-only

Get an extract job's status and, once completed, the extracted data. Returns the job object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `extract_job_id` (`string`, required): ID of the extract job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"extract_job_id":"<extract_job_id>"}`

### `hyperbrowser_extract_status_get`

Get Extract Job Status · Read-only

Get only the status of an extract job. Returns the status. Use hyperbrowser_extract_get for the data.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `extract_job_id` (`string`, required): ID of the extract job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"extract_job_id":"<extract_job_id>"}`

### `hyperbrowser_gemini_computer_use_task_get`

Get Gemini Computer Use Task · Read-only

Get a Gemini Computer Use task's status, steps and final result. Returns the task object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_gemini_computer_use_task_status_get`

Get Gemini Computer Use Task Status · Read-only

Get only the status of a Gemini Computer Use task. Returns the status. Use hyperbrowser_gemini_computer_use_task_get for the result.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_hyper_agent_task_get`

Get HyperAgent Task · Read-only

Get a HyperAgent task's status, steps and final result. Returns the task object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_hyper_agent_task_status_get`

Get HyperAgent Task Status · Read-only

Get only the status of a HyperAgent task. Returns the status. Use hyperbrowser_hyper_agent_task_get for the result.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_openai_cua_task_get`

Get OpenAI CUA Task · Read-only

Get a OpenAI CUA task's status, steps and final result. Returns the task object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_openai_cua_task_status_get`

Get OpenAI CUA Task Status · Read-only

Get only the status of a OpenAI CUA task. Returns the status. Use hyperbrowser_openai_cua_task_get for the result.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_profile_get`

Get Profile · Read-only

Get one browser profile by ID. Returns the profile with its name and timestamps. Get profile IDs from hyperbrowser_profiles_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `profile_id` (`string`, required): ID of the profile. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"profile_id":"<profile_id>"}`

### `hyperbrowser_profiles_list`

List Profiles · Read-only

List saved browser profiles, optionally filtered by name, a page at a time. Returns profile IDs and names.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `limit` (`number`, optional): Number of results per page. Example: 20.
- `name` (`string`, optional): Filter by profile name. Example: checkout-login.
- `page` (`number`, optional): Page number of results, starting at 1. Example: 1.

Example input: `{}`

### `hyperbrowser_sandbox_get`

Get Sandbox · Read-only

Get one sandbox by ID with its status, image, resources and exposed ports. Returns the sandbox object. Get IDs from hyperbrowser_sandboxes_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `sandbox_id` (`string`, required): ID of the sandbox. Example: sbx_9a8b7c6d.

Example input: `{"sandbox_id":"<sandbox_id>"}`

### `hyperbrowser_sandboxes_list`

List Sandboxes · Read-only

List sandboxes, optionally filtered by status, creation time or ID search, a page at a time. Returns sandbox IDs, statuses and images.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `end` (`number`, optional): Only sandboxes created at or before this time, in epoch milliseconds. Example: 1767312000000.
- `limit` (`integer`, optional): Number of results per page. Example: 20.
- `page` (`integer`, optional): Page number of results, starting at 1. Example: 1.
- `search` (`string`, optional): Search text. Example: warm.
- `start` (`number`, optional): Only sandboxes created at or after this time, in epoch milliseconds. Example: 1767225600000.
- `status` (`string`, optional): Filter by status. Example: active. One of: `active`, `stopping`, `closed`, `close-error`, `error`.

Example input: `{}`

### `hyperbrowser_scrape_batch_get`

Get Batch Scrape Job · Read-only

Get a batch scrape job's status and the scraped pages, a page of results at a time. Returns the job with per-URL results. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `batch_scrape_job_id` (`string`, required): ID of the batch scrape job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"batch_scrape_job_id":"<batch_scrape_job_id>"}`

### `hyperbrowser_scrape_batch_status_get`

Get Batch Scrape Job Status · Read-only

Get only the status of a batch scrape job. Returns the status. Use hyperbrowser_scrape_batch_get for the results.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `batch_scrape_job_id` (`string`, required): ID of the batch scrape job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"batch_scrape_job_id":"<batch_scrape_job_id>"}`

### `hyperbrowser_scrape_get`

Get Scrape Job · Read-only

Get a scrape job's status and, once completed, the page content in the requested formats. Returns the job object. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `scrape_job_id` (`string`, required): ID of the scrape job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"scrape_job_id":"<scrape_job_id>"}`

### `hyperbrowser_scrape_status_get`

Get Scrape Job Status · Read-only

Get only the status of a scrape job (pending, running, completed or failed). Returns the status. Use hyperbrowser_scrape_get for the content.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `scrape_job_id` (`string`, required): ID of the scrape job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"scrape_job_id":"<scrape_job_id>"}`

### `hyperbrowser_session_downloads_url_get`

Get Session Downloads URL · Read-only

Get a URL to the archive of files downloaded during a session started with save_downloads. Returns the archive status and a temporary URL.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"session_id":"<session_id>"}`

### `hyperbrowser_session_get`

Get Browser Session · Read-only

Get one browser session by ID with its status, settings, start and end times and connection URLs. Returns the session object. Get session IDs from hyperbrowser_sessions_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"session_id":"<session_id>"}`

### `hyperbrowser_session_recording_url_get`

Get Session Recording URL · Read-only

Get the replay (rrweb) recording URL of a session started with web recording enabled. Returns the recording status and a temporary URL.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"session_id":"<session_id>"}`

### `hyperbrowser_session_video_recording_url_get`

Get Session Video Recording URL · Read-only

Get the video recording URL of a session started with video recording enabled. Returns the recording status and a temporary URL.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"session_id":"<session_id>"}`

### `hyperbrowser_session_webmcp_tools_list`

List Session WebMCP Tools · Read-only

List the WebMCP tools that pages open in a session expose, across tabs and frames. Returns tool references, names, input schemas and their source page. The session must have WebMCP enabled. Call a tool with hyperbrowser_session_webmcp_tool_invoke.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"session_id":"<session_id>"}`

### `hyperbrowser_sessions_list`

List Browser Sessions · Read-only

List browser sessions, optionally only active, closed or errored ones, a page at a time. Returns session IDs, statuses and times. Use hyperbrowser_session_get for one session.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page` (`number`, optional): Page number of results, starting at 1. Example: 1.
- `status` (`string`, optional): Filter by status. Example: active. One of: `active`, `closed`, `error`, `close-error`.

Example input: `{}`

### `hyperbrowser_volume_get`

Get Volume · Read-only

Get one volume by ID. Returns the volume with its name, size and timestamps.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `volume_id` (`string`, required): ID of the volume. Example: vol_12ab34.

Example input: `{"volume_id":"<volume_id>"}`

### `hyperbrowser_volumes_list`

List Volumes · Read-only

List persistent volumes, optionally searched by name or ID, a page at a time. Returns volume IDs, names and sizes.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `limit` (`integer`, optional): Number of results per page. Example: 20.
- `page` (`integer`, optional): Page number of results, starting at 1. Example: 1.
- `search` (`string`, optional): Search text. Example: warm.

Example input: `{}`

### `hyperbrowser_web_crawl_get`

Get Web Crawl Job · Read-only

Get a web crawl job's status and page results, a page at a time. Returns the job with results. Poll until the status is completed or failed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `web_crawl_job_id` (`string`, required): ID of the web crawl job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `batch_size` (`integer`, optional): Number of page results per response page. Example: 20.
- `page` (`integer`, optional): Page number of results, starting at 1. Example: 1.

Example input: `{"web_crawl_job_id":"<web_crawl_job_id>"}`

### `hyperbrowser_web_crawl_status_get`

Get Web Crawl Job Status · Read-only

Get only the status of a web crawl job. Returns the status. Use hyperbrowser_web_crawl_get for the pages.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `web_crawl_job_id` (`string`, required): ID of the web crawl job. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"web_crawl_job_id":"<web_crawl_job_id>"}`

### `hyperbrowser_web_fetch`

Fetch Web Page · Read-only

Fetch a web page in a cloud browser and wait for the content as Markdown, HTML, links, a screenshot, JSON or branding data. Returns the content directly. Use it for one page now; use hyperbrowser_scrape_start for async jobs.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `url` (`string`, required): URL of the page to start from. Example: https://example.com/pricing.
- `browser` (`object`, optional): Browser settings: screen, profileId, solveCaptchas and location. Example: {"screen": {"width": 1440, "height": 900}}
- `cache` (`object`, optional): Cache settings: maxAgeSeconds to accept a cached copy up to that age. Example: {"maxAgeSeconds": 3600}
- `navigation` (`object`, optional): Navigation settings: waitUntil (load, domcontentloaded, networkidle), timeoutMs and waitFor. Example: {"waitUntil": "load", "timeoutMs": 30000}
- `outputs` (`object`, optional): Output settings: formats (markdown, html, links, screenshot, json, branding), sanitize, includeSelectors and excludeSelectors. Example: {"formats": ["markdown"]}
- `stealth` (`string`, optional): Stealth level: none, auto or ultra. Example: auto. One of: `none`, `auto`, `ultra`.

Example input: `{"url":"<url>"}`

### `hyperbrowser_web_search`

Search the Web · Read-only

Run a web search with optional location and filters (site, file type, title or URL terms, exact phrase). Returns results with titles, URLs and descriptions. Use hyperbrowser_web_fetch to read a result.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `query` (`string`, required): Search query. Example: best browser automation tools 2026.
- `filters` (`object`, optional): Search filters: site, excludeSite, filetype, intitle, inurl, exactPhrase, excludeTerms and boostTerms. Example: {"site": "github.com"}
- `location` (`object`, optional): Location to search from: country, state and city. Example: {"country": "US"}
- `max_age_seconds` (`integer`, optional): Accept cached results up to this age in seconds. Example: 86400.
- `page` (`integer`, optional): Page number of results, starting at 1. Example: 1.

Example input: `{"query":"<query>"}`

### `hyperbrowser_browser_use_task_start`

Start Browser Use Task · Write

Start a task that the open-source Browser Use agent runs in a cloud browser from natural-language instructions. Returns a task ID and the live-view URL; get the result with hyperbrowser_browser_use_task_get. Tasks are billed per step. Other agents: hyperbrowser_hyper_agent_task_start, hyperbrowser_browser_use_task_start, hyperbrowser_claude_computer_use_task_start, hyperbrowser_openai_cua_task_start, hyperbrowser_gemini_computer_use_task_start.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `task` (`string`, required): Natural-language instructions for the agent. Example: Go to news.ycombinator.com and return the titles of the top 3 stories.
- `keep_browser_open` (`boolean`, optional): Set to true to keep the session open after the task ends. Example: false.
- `llm` (`string`, optional): Model that runs the agent. Example: gpt-4o. One of: `gpt-4o`, `gpt-4o-mini`, `gpt-4.1`, `gpt-4.1-mini`, `claude-sonnet-5`, `claude-sonnet-4-6`, `claude-haiku-5-5`, `gemini-2.0-flash`, `gemini-2.5-flash`, `gemini-4-hydrogen`, `gemini-3.8-flash`.
- `max_actions_per_step` (`number`, optional): Maximum browser actions per agent step. Example: 5.
- `max_failures` (`number`, optional): Maximum consecutive failures before the task stops. Example: 3.
- `max_steps` (`number`, optional): Maximum number of agent steps. Example: 20.
- `message_context` (`string`, optional): Extra context for the agent, such as facts it should know. Example: Prices are listed in USD.
- `session_id` (`string`, optional): ID of an existing browser session to run in, from hyperbrowser_session_create; a new session is created when empty. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}
- `use_vision` (`boolean`, optional): Set to true to let the agent use screenshots. Example: true.
- `validate_output` (`boolean`, optional): Set to true to have the agent check its own output before finishing. Example: false.
- `version` (`string`, optional): Agent version to run; defaults to the latest. Example: 0.1.40. One of: `0.1.40`, `0.7.10`, `latest`.

Example input: `{"task":"<task>"}`

### `hyperbrowser_claude_computer_use_task_start`

Start Claude Computer Use Task · Write

Start a task that Anthropic Claude computer use runs in a cloud browser from natural-language instructions. Returns a task ID and the live-view URL; get the result with hyperbrowser_claude_computer_use_task_get. Tasks are billed per step. Other agents: hyperbrowser_hyper_agent_task_start, hyperbrowser_browser_use_task_start, hyperbrowser_claude_computer_use_task_start, hyperbrowser_openai_cua_task_start, hyperbrowser_gemini_computer_use_task_start.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `task` (`string`, required): Natural-language instructions for the agent. Example: Go to news.ycombinator.com and return the titles of the top 3 stories.
- `keep_browser_open` (`boolean`, optional): Set to true to keep the session open after the task ends. Example: false.
- `llm` (`string`, optional): Model that runs the agent. Example: claude-opus-5-5. One of: `claude-opus-5-5`, `claude-fable-5-2`, `claude-fable-5-1`, `claude-opus-5`, `claude-opus-4-5`, `claude-opus-4-6`, `claude-opus-4-7`, `claude-opus-4-8`, `claude-haiku-5-5`, `claude-haiku-4-5-20251001`, `claude-sonnet-5-5`, `claude-sonnet-5`, `claude-sonnet-4-6`.
- `max_failures` (`number`, optional): Maximum consecutive failures before the task stops. Example: 3.
- `max_steps` (`number`, optional): Maximum number of agent steps. Example: 20.
- `reasoning_effort` (`string`, optional): Reasoning effort for the model. Example: medium. One of: `low`, `medium`, `high`, `xhigh`, `max`.
- `session_id` (`string`, optional): ID of an existing browser session to run in, from hyperbrowser_session_create; a new session is created when empty. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}
- `use_computer_action` (`boolean`, optional): Set to true to drive the browser through the computer action endpoint. Example: false.

Example input: `{"task":"<task>"}`

### `hyperbrowser_crawl_start`

Start Crawl Job · Write

Start crawling a website from a URL, following links up to max_pages, and scrape each page. Returns a job ID; get the pages with hyperbrowser_crawl_get. Use hyperbrowser_web_crawl_start for the newer web crawl API with fetch-style outputs.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `url` (`string`, required): URL of the page to start from. Example: https://example.com/pricing.
- `exclude_patterns` (`array`, optional): URL path patterns to skip. Example: ["/blog/*"]
- `follow_links` (`boolean`, optional): Set to true to follow links found on crawled pages. Example: true.
- `ignore_sitemap` (`boolean`, optional): Set to true to skip the site's sitemap when discovering pages. Example: false.
- `include_patterns` (`array`, optional): URL path patterns to crawl; other pages are skipped. Example: ["/docs/*"]
- `max_pages` (`integer`, optional): Maximum number of pages to crawl. Example: 10.
- `scrape_options` (`object`, optional): Scrape settings: formats (html, links, markdown, screenshot), onlyMainContent, includeTags, excludeTags, waitFor, timeout and waitUntil. Example: {"formats": ["markdown", "links"], "onlyMainContent": true}
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}

Example input: `{"url":"<url>"}`

### `hyperbrowser_extract_start`

Start Extract Job · Write

Start extracting structured data from one or more URLs with an AI model, guided by a prompt and/or a JSON Schema. Returns a job ID; get the data with hyperbrowser_extract_get. A URL ending in /* also follows links on that site.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `urls` (`array`, required): URLs of the pages to process. Example: ["https://example.com/pricing", "https://example.com/about"]
- `max_links` (`number`, optional): Maximum number of links to follow when a URL ends with /*. Example: 5.
- `prompt` (`string`, optional): What to extract from the pages. Example: Extract every plan name with its monthly price.
- `schema` (`object`, optional): JSON Schema the extracted data must follow. Example: {"type": "object", "properties": {"plans": {"type": "array", "items": {"type": "object", "properties": {"name": {"type": "string"}, "price": {"type": "string...
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}
- `system_prompt` (`string`, optional): System prompt for the extraction model. Example: You extract pricing data from SaaS websites.
- `wait_for` (`number`, optional): Milliseconds to wait after page load before extracting. Example: 2000.

Example input: `{"urls":["https://example.com/pricing","https://example.com/about"]}`

### `hyperbrowser_gemini_computer_use_task_start`

Start Gemini Computer Use Task · Write

Start a task that Google Gemini computer use runs in a cloud browser from natural-language instructions. Returns a task ID and the live-view URL; get the result with hyperbrowser_gemini_computer_use_task_get. Tasks are billed per step. Other agents: hyperbrowser_hyper_agent_task_start, hyperbrowser_browser_use_task_start, hyperbrowser_claude_computer_use_task_start, hyperbrowser_openai_cua_task_start, hyperbrowser_gemini_computer_use_task_start.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `task` (`string`, required): Natural-language instructions for the agent. Example: Go to news.ycombinator.com and return the titles of the top 3 stories.
- `keep_browser_open` (`boolean`, optional): Set to true to keep the session open after the task ends. Example: false.
- `llm` (`string`, optional): Model that runs the agent. Example: gemini-4-hydrogen. One of: `gemini-4-hydrogen`, `gemini-4-carbon`, `gemini-4-argon`, `gemini-3.8-flash`, `gemini-3.7-flash`, `gemini-3.6-flash`, `gemini-3.5-flash`, `gemini-3.5-flash-lite`, `gemini-3-flash-preview`, `gemini-2.5-computer-use-preview-10-2025`.
- `max_failures` (`number`, optional): Maximum consecutive failures before the task stops. Example: 3.
- `max_steps` (`number`, optional): Maximum number of agent steps. Example: 20.
- `reasoning_effort` (`string`, optional): Reasoning effort for the model. Example: medium. One of: `minimal`, `low`, `medium`, `high`, `xhigh`.
- `session_id` (`string`, optional): ID of an existing browser session to run in, from hyperbrowser_session_create; a new session is created when empty. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}
- `use_computer_action` (`boolean`, optional): Set to true to drive the browser through the computer action endpoint. Example: false.

Example input: `{"task":"<task>"}`

### `hyperbrowser_hyper_agent_task_start`

Start HyperAgent Task · Write

Start a task that Hyperbrowser's own browser agent runs in a cloud browser from natural-language instructions. Returns a task ID and the live-view URL; get the result with hyperbrowser_hyper_agent_task_get. Tasks are billed per step. Other agents: hyperbrowser_hyper_agent_task_start, hyperbrowser_browser_use_task_start, hyperbrowser_claude_computer_use_task_start, hyperbrowser_openai_cua_task_start, hyperbrowser_gemini_computer_use_task_start.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `task` (`string`, required): Natural-language instructions for the agent. Example: Go to news.ycombinator.com and return the titles of the top 3 stories.
- `keep_browser_open` (`boolean`, optional): Set to true to keep the session open after the task ends. Example: false.
- `llm` (`string`, optional): Model that runs the agent. Example: gpt-5.5. One of: `gpt-6.1-terra`, `gpt-6-luna`, `gpt-5.6-luna`, `gpt-5.5`, `gpt-5.2`, `gpt-5.1`, `gpt-5`, `gpt-5-mini`, `gpt-4o`, `gpt-4o-mini`, `gpt-4.1`, `gpt-4.1-mini`, `claude-sonnet-5`, `claude-sonnet-4-6`, `claude-haiku-5-5`, `gemini-2.5-flash`, `gemini-4-hydrogen`, `gemini-3.8-flash`, `gemini-3.7-flash`, `gemini-3-flash-preview`.
- `max_steps` (`number`, optional): Maximum number of agent steps. Example: 20.
- `session_id` (`string`, optional): ID of an existing browser session to run in, from hyperbrowser_session_create; a new session is created when empty. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}
- `version` (`string`, optional): Agent version to run; defaults to the latest. Example: 1.1.0. One of: `0.8.0`, `1.1.0`.

Example input: `{"task":"<task>"}`

### `hyperbrowser_openai_cua_task_start`

Start OpenAI CUA Task · Write

Start a task that OpenAI's computer-using agent (CUA) runs in a cloud browser from natural-language instructions. Returns a task ID and the live-view URL; get the result with hyperbrowser_openai_cua_task_get. Tasks are billed per step. Other agents: hyperbrowser_hyper_agent_task_start, hyperbrowser_browser_use_task_start, hyperbrowser_claude_computer_use_task_start, hyperbrowser_openai_cua_task_start, hyperbrowser_gemini_computer_use_task_start.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `task` (`string`, required): Natural-language instructions for the agent. Example: Go to news.ycombinator.com and return the titles of the top 3 stories.
- `keep_browser_open` (`boolean`, optional): Set to true to keep the session open after the task ends. Example: false.
- `llm` (`string`, optional): Model that runs the agent. Example: gpt-5.5. One of: `computer-use-preview`, `gpt-6-astra`, `gpt-6.1-sol`, `gpt-6.1-terra`, `gpt-6-sol`, `gpt-6-luna`, `gpt-5.6-sol`, `gpt-5.6-terra`, `gpt-5.6-luna`, `gpt-5.5`, `gpt-5.4`, `gpt-5.4-mini`.
- `max_failures` (`number`, optional): Maximum consecutive failures before the task stops. Example: 3.
- `max_steps` (`number`, optional): Maximum number of agent steps. Example: 20.
- `reasoning_effort` (`string`, optional): Reasoning effort for the model. Example: medium. One of: `none`, `low`, `medium`, `high`, `xhigh`, `max`.
- `session_id` (`string`, optional): ID of an existing browser session to run in, from hyperbrowser_session_create; a new session is created when empty. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}
- `use_computer_action` (`boolean`, optional): Set to true to drive the browser through the computer action endpoint. Example: false.

Example input: `{"task":"<task>"}`

### `hyperbrowser_profile_create`

Create Profile · Write

Create a browser profile that stores cookies, local storage and cache between sessions. Returns the profile ID. Load it with the profile option of hyperbrowser_session_create.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `name` (`string`, optional): Name of the profile. Example: checkout-login.

Example input: `{}`

### `hyperbrowser_profile_fork`

Fork Profile · Write

Copy an existing browser profile into a new one. Returns the new profile ID. Use it to branch a logged-in state without changing the original.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `profile_id` (`string`, required): ID of the profile to copy. Example: 5b1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `name` (`string`, optional): Name of the new profile. Example: checkout-login.

Example input: `{"profile_id":"<profile_id>"}`

### `hyperbrowser_sandbox_create`

Create Sandbox · Write

Start a CPU sandbox (a cloud VM) from an image such as node or python, or from a memory snapshot, with optional CPU, memory, disk, ports, volume mounts and outbound network rules. Returns the sandbox ID and connection details. Stop it with hyperbrowser_sandbox_stop when done.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `allow_internet_access` (`boolean`, optional): Outbound internet access policy. Example: true.
- `allow_out` (`array`, optional): Outbound allow rules (IPv4 address, CIDR range, or domain). macOS supports IPv4 addresses and CIDRs only. Example: ["api.github.com"]
- `deny_out` (`array`, optional): Outbound deny rules (IPv4 address or CIDR range). macOS supports IPv4 addresses and CIDRs only. Example: ["10.0.0.0/8"]
- `disk_size_mi_b` (`integer`, optional): Disk size in MiB for image launches. GPU sandboxes default to and require at least 12288 MiB (12 GiB), with a maximum of 65536 MiB (64 GiB), subject to plan and host limits. GPU disk is ephemeral and files are deleted when the sandbox stops. For macOS, omit this field to use the source disk size; resizing is not supported. Example: 8192.
- `enable_recording` (`boolean`, optional): Enable recording. For macOS, captures an HLS desktop recording available in the dashboard Recording tab. Example: false.
- `exposed_ports` (`array`, optional): Ports to expose at launch. Example: [{"port": 8080}]
- `image_id` (`string`, optional): Pin a specific image revision. Requires imageName. Example: img_7f3a2b.
- `image_name` (`string`, optional): Image name to launch from. CPU sandboxes can use a base image such as node or python, or a custom image. GPU sandboxes require a GPU-ready image built by your team; wait for gpuImageReady on its build. For macOS, use macos-base with runtimeClass macos; each image launch creates a new writable disk. Example: python.
- `mem_mi_b` (`integer`, optional): System memory (RAM) in MiB for image launches or macOS disk launches. For GPU sandboxes this is separate from GPU memory (VRAM), which is determined by gpu. macOS requires at least 4096 MiB, in multiples of 1024 MiB. Plan and host limits apply. Example: 2048.
- `mounts` (`object`, optional): Volume mounts, keyed by absolute mount path inside a CPU sandbox. GPU and macOS sandboxes do not support shared volume mounts. Example: {"/data": {"volumeId": "vol_12ab34"}}
- `region` (`string`, optional): Region to run in. Example: us. One of: `us`, `us-central`, `us-west`, `us-east`, `asia-south`, `europe-west`.
- `runtime_class` (`string`, optional): Runtime selection. Requests without GPU or macOS options create CPU sandboxes (firecracker). Setting gpu selects gvisor-cpu automatically; omit runtimeClass for GPU creation. Set macos explicitly for a Mac sandbox. Example: firecracker. One of: `firecracker`, `gvisor-cpu`, `macos`.
- `snapshot_id` (`string`, optional): Pin a specific snapshot revision. Requires snapshotName. Example: snap_4c9d1e.
- `snapshot_name` (`string`, optional): CPU sandbox memory snapshot to restore. GPU sandboxes do not support snapshots; macOS uses diskId instead. Example: warm-python.
- `timeout_minutes` (`integer`, optional): Maximum lifetime in minutes. Example: 30.
- `vcpus` (`integer`, optional): vCPU count for image launches or macOS disk launches. macOS requires at least 2 vCPUs; available sizes depend on your plan and host capacity. Example: 2.

Example input: `{}`

### `hyperbrowser_scrape_batch_start`

Start Batch Scrape Job · Write

Start scraping a list of URLs in parallel. Returns a job ID; get the results with hyperbrowser_scrape_batch_get. Use hyperbrowser_scrape_start for a single URL.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `urls` (`array`, required): URLs of the pages to process. Example: ["https://example.com/pricing", "https://example.com/about"]
- `scrape_options` (`object`, optional): Scrape settings: formats (html, links, markdown, screenshot), onlyMainContent, includeTags, excludeTags, waitFor, timeout and waitUntil. Example: {"formats": ["markdown", "links"], "onlyMainContent": true}
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}

Example input: `{"urls":["https://example.com/pricing","https://example.com/about"]}`

### `hyperbrowser_scrape_start`

Start Scrape Job · Write

Start scraping one URL in a cloud browser and return its content as Markdown, HTML, links or a screenshot. Returns a job ID; get the result with hyperbrowser_scrape_get. Use hyperbrowser_web_fetch for a synchronous fetch and hyperbrowser_scrape_batch_start for many URLs.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `url` (`string`, required): URL of the page to start from. Example: https://example.com/pricing.
- `scrape_options` (`object`, optional): Scrape settings: formats (html, links, markdown, screenshot), onlyMainContent, includeTags, excludeTags, waitFor, timeout and waitUntil. Example: {"formats": ["markdown", "links"], "onlyMainContent": true}
- `session_options` (`object`, optional): Browser session settings for the job, such as useStealth, useProxy, proxyCountry, solveCaptchas, adblock, screen and profile (same fields as hyperbrowser_session_create). Example: {"useStealth": true, "solveCaptchas": true, "adblock": true}

Example input: `{"url":"<url>"}`

### `hyperbrowser_session_create`

Create Browser Session · Write

Start a cloud browser session with optional stealth, proxy, CAPTCHA solving, ad blocking, fingerprint, profile and recording settings. Returns the session ID, live-view URL and the CDP/WebSocket endpoint to drive it. Stop it with hyperbrowser_session_stop when done, since sessions are billed while open. Use the scrape, extract or task tools instead when you only need a result.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `accept_cookies` (`boolean`, optional): Set to true to accept cookie banners automatically. Example: true.
- `adblock` (`boolean`, optional): Set to true to block ads. Example: true.
- `annoyances` (`boolean`, optional): Set to true to block cookie banners and pop-ups. Example: true.
- `device` (`array`, optional): Device types to fingerprint as: desktop or mobile. Example: ["desktop"]
- `enable_video_web_recording` (`boolean`, optional): Set to true to also record video; enable_web_recording must be true. Example: false.
- `enable_web_mcp` (`boolean`, optional): Set to true to discover and invoke WebMCP tools on pages. Example: false.
- `enable_web_recording` (`boolean`, optional): Set to true to record the session for replay. Example: true.
- `extension_ids` (`array`, optional): IDs of uploaded extensions to load, from hyperbrowser_extensions_list. Example: ["a1b2c3d4-e5f6-4a7b-8c9d-0e1f2a3b4c5d"]
- `locales` (`array`, optional): Browser locales, for example en. Example: ["en"]
- `operating_systems` (`array`, optional): Operating systems to fingerprint as: windows, android, macos, linux or ios. Example: ["macos"]
- `platform` (`array`, optional): Browsers to fingerprint as, for example chrome, firefox, safari or edge. Example: ["chrome"]
- `profile` (`object`, optional): Saved profile to load: id, plus persistChanges to save cookies and storage back to it. Example: {"id": "5b1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e", "persistChanges": true}
- `proxy_city` (`string`, optional): City for the proxy, for example new york; cannot be combined with proxy_state. Example: new york.
- `proxy_country` (`string`, optional): Two-letter country code for the proxy. Example: US. One of: `AD`, `AE`, `AF`, `AL`, `AM`, `AO`, `AR`, `AT`, `AU`, `AW`, `AZ`, `BA`, `BD`, `BE`, `BG`, `BH`, `BJ`, `BO`, `BR`, `BS`, `BT`, `BY`, `BZ`, `CA`, `CF`, `CH`, `CI`, `CL`, `CM`, `CN`, `CO`, `CR`, `CU`, `CY`, `CZ`, `DE`, `DJ`, `DK`, `DM`, `EC`, `EE`, `EG`, `ES`, `ET`, `EU`, `FI`, `FJ`, `FR`, `GB`, `GE`, `GH`, `GM`, `GR`, `HK`, `HN`, `HR`, `HT`, `HU`, `ID`, `IE`, `IL`, `IN`, `IQ`, `IR`, `IS`, `IT`, `JM`, `JO`, `JP`, `KE`, `KH`, `KR`, `KW`, `KZ`, `LB`, `LI`, `LR`, `LT`, `LU`, `LV`, `MA`, `MC`, `MD`, `ME`, `MG`, `MK`, `ML`, `MM`, `MN`, `MR`, `MT`, `MU`, `MV`, `MX`, `MY`, `MZ`, `NG`, `NL`, `NO`, `NZ`, `OM`, `PA`, `PE`, `PH`, `PK`, `PL`, `PR`, `PT`, `PY`, `QA`, `RANDOM_COUNTRY`, `RO`, `RS`, `RU`, `SA`, `SC`, `SD`, `SE`, `SG`, `SI`, `SK`, `SN`, `SS`, `TD`, `TG`, `TH`, `TM`, `TN`, `TR`, `TT`, `TW`, `UA`, `UG`, `US`, `UY`, `UZ`, `VE`, `VG`, `VN`, `YE`, `ZA`, `ZM`, `ZW`, `ad`, `ae`, `af`, `al`, `am`, `ao`, `ar`, `at`, `au`, `aw`, `az`, `ba`, `bd`, `be`, `bg`, `bh`, `bj`, `bo`, `br`, `bs`, `bt`, `by`, `bz`, `ca`, `cf`, `ch`, `ci`, `cl`, `cm`, `cn`, `co`, `cr`, `cu`, `cy`, `cz`, `de`, `dj`, `dk`, `dm`, `ec`, `ee`, `eg`, `es`, `et`, `eu`, `fi`, `fj`, `fr`, `gb`, `ge`, `gh`, `gm`, `gr`, `hk`, `hn`, `hr`, `ht`, `hu`, `id`, `ie`, `il`, `in`, `iq`, `ir`, `is`, `it`, `jm`, `jo`, `jp`, `ke`, `kh`, `kr`, `kw`, `kz`, `lb`, `li`, `lr`, `lt`, `lu`, `lv`, `ma`, `mc`, `md`, `me`, `mg`, `mk`, `ml`, `mm`, `mn`, `mr`, `mt`, `mu`, `mv`, `mx`, `my`, `mz`, `ng`, `nl`, `no`, `nz`, `om`, `pa`, `pe`, `ph`, `pk`, `pl`, `pr`, `pt`, `py`, `qa`, `ro`, `rs`, `ru`, `sa`, `sc`, `sd`, `se`, `sg`, `si`, `sk`, `sn`, `ss`, `td`, `tg`, `th`, `tm`, `tn`, `tr`, `tt`, `tw`, `ua`, `ug`, `us`, `uy`, `uz`, `ve`, `vg`, `vn`, `ye`, `za`, `zm`, `zw`.
- `proxy_state` (`string`, optional): US state code for the proxy; cannot be combined with proxy_city. Example: CA. One of: `AL`, `AK`, `AZ`, `AR`, `CA`, `CO`, `CT`, `DE`, `FL`, `GA`, `HI`, `ID`, `IL`, `IN`, `IA`, `KS`, `KY`, `LA`, `ME`, `MD`, `MA`, `MI`, `MN`, `MS`, `MO`, `MT`, `NE`, `NV`, `NH`, `NJ`, `NM`, `NY`, `NC`, `ND`, `OH`, `OK`, `OR`, `PA`, `RI`, `SC`, `SD`, `TN`, `TX`, `UT`, `VT`, `VA`, `WA`, `WV`, `WI`, `WY`, `al`, `ak`, `az`, `ar`, `ca`, `co`, `ct`, `de`, `fl`, `ga`, `hi`, `id`, `il`, `in`, `ia`, `ks`, `ky`, `la`, `me`, `md`, `ma`, `mi`, `mn`, `ms`, `mo`, `mt`, `ne`, `nv`, `nh`, `nj`, `nm`, `ny`, `nc`, `nd`, `oh`, `ok`, `or`, `pa`, `ri`, `sc`, `sd`, `tn`, `tx`, `ut`, `vt`, `va`, `wa`, `wv`, `wi`, `wy`.
- `region` (`string`, optional): Region to run in. Example: us. One of: `us`, `us-central`, `us-west`, `us-east`, `asia-south`, `europe-west`.
- `save_downloads` (`boolean`, optional): Set to true to keep files downloaded during the session. Example: false.
- `screen` (`object`, optional): Screen size as width and height in pixels. Example: {"width": 1440, "height": 900}
- `solve_captchas` (`boolean`, optional): Set to true to solve CAPTCHAs automatically (paid plans). Example: true.
- `static_ip_id` (`string`, optional): ID of a static IP to use. Example: sip_123.
- `timeout_minutes` (`number`, optional): Maximum lifetime in minutes. Example: 30.
- `trackers` (`boolean`, optional): Set to true to block trackers. Example: true.
- `url_blocklist` (`array`, optional): URL patterns the browser must not load. Example: ["*.doubleclick.net"]
- `use_proxy` (`boolean`, optional): Set to true to route traffic through a residential proxy. Example: false.
- `use_stealth` (`boolean`, optional): Set to true to use stealth mode against bot detection. Example: true.
- `use_ultra_stealth` (`boolean`, optional): Set to true to use ultra stealth mode (paid plans). Example: false.

Example input: `{}`

### `hyperbrowser_session_webmcp_tool_invoke`

Invoke Session WebMCP Tool · Write

Invoke a WebMCP tool exposed by a page in a session and wait for the result. Returns the tool output, or awaiting_submission with an invocation ID for forms. Do not retry blindly if the response is lost, since the tool may have run. Get tool references from hyperbrowser_session_webmcp_tools_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.
- `tool_ref` (`string`, required): Tool reference from hyperbrowser_session_webmcp_tools_list. Example: tool_ref_abc123.
- `input` (`object`, optional): Arguments matching the WebMCP tool's input schema. Example: {"query": "blue running shoes"}
- `timeout_seconds` (`integer`, optional): Tool execution budget in seconds, 1 to 120. Example: 60.

Example input: `{"session_id":"<session_id>","tool_ref":"<tool_ref>"}`

### `hyperbrowser_volume_create`

Create Volume · Write

Create a persistent volume that can be mounted into sandboxes. Returns the volume ID. Mount it with the mounts option of hyperbrowser_sandbox_create.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `name` (`string`, required): Name of the volume. Example: checkout-login.

Example input: `{"name":"<name>"}`

### `hyperbrowser_web_crawl_start`

Start Web Crawl Job · Write

Start an async crawl that follows links from a URL and returns each page in the requested output formats. Returns a job ID; get the pages with hyperbrowser_web_crawl_get.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `url` (`string`, required): URL of the page to start from. Example: https://example.com/pricing.
- `browser` (`object`, optional): Browser settings: screen, profileId, solveCaptchas and location. Example: {"screen": {"width": 1440, "height": 900}}
- `cache` (`object`, optional): Cache settings: maxAgeSeconds to accept a cached copy up to that age. Example: {"maxAgeSeconds": 3600}
- `crawl_options` (`object`, optional): Crawl settings: maxPages, followLinks, ignoreSitemap, includePatterns and excludePatterns. Example: {"maxPages": 10, "followLinks": true}
- `navigation` (`object`, optional): Navigation settings: waitUntil (load, domcontentloaded, networkidle), timeoutMs and waitFor. Example: {"waitUntil": "load", "timeoutMs": 30000}
- `outputs` (`object`, optional): Output settings: formats (markdown, html, links, screenshot, json, branding), sanitize, includeSelectors and excludeSelectors. Example: {"formats": ["markdown"]}
- `stealth` (`string`, optional): Stealth level: none, auto or ultra. Example: auto. One of: `none`, `auto`, `ultra`.

Example input: `{"url":"<url>"}`

### `hyperbrowser_browser_use_task_stop`

Stop Browser Use Task · Destructive

Stop a running Browser Use task. Returns the result. The partial result stays available from hyperbrowser_browser_use_task_get.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_claude_computer_use_task_stop`

Stop Claude Computer Use Task · Destructive

Stop a running Claude Computer Use task. Returns the result. The partial result stays available from hyperbrowser_claude_computer_use_task_get.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_gemini_computer_use_task_stop`

Stop Gemini Computer Use Task · Destructive

Stop a running Gemini Computer Use task. Returns the result. The partial result stays available from hyperbrowser_gemini_computer_use_task_get.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_hyper_agent_task_stop`

Stop HyperAgent Task · Destructive

Stop a running HyperAgent task. Returns the result. The partial result stays available from hyperbrowser_hyper_agent_task_get.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_openai_cua_task_stop`

Stop OpenAI CUA Task · Destructive

Stop a running OpenAI CUA task. Returns the result. The partial result stays available from hyperbrowser_openai_cua_task_get.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `task_id` (`string`, required): ID of the task. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"task_id":"<task_id>"}`

### `hyperbrowser_profile_delete`

Delete Profile · Destructive

Delete a browser profile and its stored data. Returns the result.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `profile_id` (`string`, required): ID of the profile. Example: 5b1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.

Example input: `{"profile_id":"<profile_id>"}`

### `hyperbrowser_sandbox_stop`

Stop Sandbox · Destructive

Stop a running sandbox and release its VM. Returns the result.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `sandbox_id` (`string`, required): ID of the sandbox. Example: sbx_9a8b7c6d.

Example input: `{"sandbox_id":"<sandbox_id>"}`

### `hyperbrowser_session_stop`

Stop Browser Session · Destructive

Stop a running browser session and release its browser. Returns the result. Recordings and downloads stay available afterwards.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `session_id` (`string`, required): ID of the session. Example: 8f2c1a7e-4b3d-4e8a-9c21-5d6f7a8b9c0d.

Example input: `{"session_id":"<session_id>"}`

### `hyperbrowser_volume_delete`

Delete Volume · Destructive

Delete a volume and its data by ID or name. Returns the result. A name shared by several volumes returns an error; use the ID then.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `volume` (`string`, required): Volume ID or name. Example: vol_12ab34.

Example input: `{"volume":"<volume>"}`

## Related

Other Automation connectors ([all 91](/agentkit/connectors/?category=automation)):

- [Mailchimp](/agentkit/connectors/mailchimp/): Scalekit connector, OAuth, 73 tools
- [Twilio](/agentkit/connectors/twilio/): Scalekit connector, Username and password, 61 tools
- [Stripe](/agentkit/connectors/stripe/): Scalekit connector, Access token, 178 tools


---

## More Scalekit documentation

| Resource | What it contains | When to use it |
|----------|-----------------|----------------|
| [/llms.txt](/llms.txt) | Structured index with routing hints per product area | Start here — find which documentation set covers your topic before loading full content |
| [/llms-full.txt](/llms-full.txt) | Complete documentation for all Scalekit products in one file | Use when you need exhaustive context across multiple products or when the topic spans several areas |
| [sitemap-0.xml](https://docs.scalekit.com/sitemap-0.xml) | Full URL list of every documentation page | Use to discover specific page URLs you can fetch for targeted, page-level answers |
