# BrowserAct — how to use (mcp.ai)

Connect your BrowserAct account and use 11 tools for web scraping straight from your AI agent. Connect with your own API key. BrowserAct runs hosted browser-automation workflows and exposes task execution, workflow discovery, templates, and proxy-region APIs.

## Option A — via MCP (recommended)
Remote MCP endpoint (HTTP, streamable): `https://api.mcp.ai/p_browseract?ms=1787291460000`
Add it as a custom/remote MCP connector in your client (Claude, Cursor, VS Code…), then authenticate when prompted. Once connected, ask the agent to use the server's tools (e.g. `browseract_cancel_task`).

## Option B — via direct REST API
Base URL: `https://api.mcp.ai/api/browseract`
Auth: `Authorization: Bearer sk_live_…` — create a workspace API key at https://mcp.ai/settings/api-keys
Discover endpoints: `GET https://api.mcp.ai/api/browseract/_endpoints`

### Endpoints
- `POST https://api.mcp.ai/api/browseract/cancel/task` — Immediately and permanently cancel a running BrowserAct task. The canceled run cannot be resumed or recovered. A task already in a terminal state is also treated as successfully canceled.
  - body: { task_id: string, business_code?: integer }
- `POST https://api.mcp.ai/api/browseract/get/task` — Return either a task's current status or its full progress, outputs, files, live-view details, and failure information.
  - body: { task_id: string, detail_level?: string }
- `POST https://api.mcp.ai/api/browseract/get/workflow` — Return a custom workflow's metadata and expected input parameter definitions before a run.
  - body: { workflow_id: string }
- `POST https://api.mcp.ai/api/browseract/get/workflow/template` — Return an official template's metadata and expected input parameter definitions before a run.
  - body: { include_description?: boolean, workflow_template_id: string }
- `POST https://api.mcp.ai/api/browseract/list/proxy/regions` — Return the supported proxy region names and codes for official template runs.
- `POST https://api.mcp.ai/api/browseract/list/tasks` — Return one cursor-controlled page of workflow tasks, optionally filtered by status, workflow, name, or creation time.
  - body: { limit?: integer, status?: string, next_cursor?: string, workflow_id?: string, created_at_to?: string, workflow_name?: string, created_at_from?: string }
- `POST https://api.mcp.ai/api/browseract/list/workflow/templates` — Search official BrowserAct workflow templates and return one cursor-controlled page of summaries.
  - body: { limit?: integer, keyword?: string, next_cursor?: string }
- `POST https://api.mcp.ai/api/browseract/list/workflows` — Return one cursor-controlled page of custom workflows in the connected workspace.
  - body: { limit?: integer, next_cursor?: string }
- `POST https://api.mcp.ai/api/browseract/resume/task` — Resume a paused task from its current state. Finished, failed, and canceled tasks cannot be resumed. BrowserAct also treats an already-running task as a successful request.
  - body: { task_id: string }
- `POST https://api.mcp.ai/api/browseract/run/workflow` — Start a custom BrowserAct workflow. This consumes credits and executes its configured browser actions, which may irreversibly modify external websites.
  - body: { profile_id?: string, workflow_id: string, callback_url?: string, input_parameters?: object[], save_browser_data?: boolean, status_change_callback_url?: string }
- `POST https://api.mcp.ai/api/browseract/run/workflow/template` — Start an official BrowserAct workflow template. This consumes credits and executes its browser actions, which may cause irreversible effects on external websites.
  - body: { callback_url?: string, proxy_region?: string, input_parameters?: object[], workflow_template_id: string, status_change_callback_url?: string }

## Example prompts
- "What can I do in BrowserAct?"
- "Show me a summary of my BrowserAct account"

## More
- Page: https://mcp.ai/browseract
- Agent spec (llms.txt): https://mcp.ai/browseract/llms.txt
- Postman collection: https://mcp.ai/browseract/postman.json