# Get live web data with BrowserAct MCP.

Use BrowserAct’s web scraping and automation tools to access live web data directly from your AI client.

Copy to AI Agent [View setup guide](/content/integrations/mcp#client-setup/index.html)

## Example workflow

1. Find a Bot  
2. Run a task  
3. Get results

AI assistant  
I'll run this Bot and check its progress.

`run_bot`Product Price Tracker

Done

`get_task_status`Collecting product prices…

Running

## MCP server URL

Use this endpoint for manual setup.

`https://mcp.browseract.com/mcp/`

**API key required** Add your BrowserAct API key to your AI client's configuration.

[Get API key](/content/reception/integrations?open-api-keys=1/index.html)

## Client setup

Choose your client and follow its setup guide.

1. 1

### Get your API key

Create or copy your BrowserAct API key.

2. 2

### Add BrowserAct to your client

Follow the configuration instructions for your client.

3. 3

### Verify the connection

Ask your AI agent to list BrowserAct tools.

---

Connect the official BrowserAct MCP server to the AI client I am using so I can access live web data through my published Bots and BrowserAct’s official templates.

### MCP server

- **Name:**`browseract`  
- **URL:**`https://mcp.browseract.com/mcp/`  
- **Transport:** Streamable HTTP  
- **Authentication:**`Authorization: Bearer <BROWSERACT_API_KEY>`

Keep the server URL exactly as shown. Use the `Authorization: Bearer` header for every request. Do not use `x-api-key`.

### 1. Check the connection

- Check whether BrowserAct tools are already available in this session. If they are, skip setup.  
- If BrowserAct is configured but unavailable, diagnose and reconnect it instead of adding a duplicate.  
- Do not call BrowserAct tools until the connection is verified.

### 2. Configure your AI client

**API key required**  
If no valid API key is configured, first show me this clickable link:

[Get your BrowserAct API key](/content/reception/integrations?open-api-keys=1/index.html)

Tell me to sign in, then create or copy a key in the **API Keys** section on that page. If a valid key is already configured, reuse it.

- Have me replace `<BROWSERACT_API_KEY>` with my real API key directly in the client’s local configuration or secure credential field.  
- Never ask me to paste my API key into this chat.  
- If the client uses an environment variable, have me set it locally for the client process and confirm it is available without printing its value.  
- Do not connect with a literal key placeholder or store a real key in a shared project file.

Connecting to the official server does not require creating a custom MCP server or publishing a new Bot first. Once connected, I can use my existing published Bots or discover official templates.

**Client setup**

- Identify the AI client, installed version and intended configuration scope. If you cannot identify the client, ask me which client I am using.  
- Consult that client’s official MCP setup documentation and prefer its native remote-server setup flow.  
- Configure `browseract` using Streamable HTTP, the server URL above and the `Authorization` Bearer header. Preserve existing MCP servers.  
- Use the configuration fields and environment-variable syntax documented for this client. Do not assume all clients accept the same JSON format or guess configuration paths.  
- If you can edit the configuration or run the supported setup commands, help me configure the connection. Otherwise, show the exact manual steps.  
- If this client does not support the required transport or authentication headers, explain the limitation and suggest a supported alternative before proceeding.

### 3. Verify tool availability

- Ask me to reload or restart the client only if needed.  
- Confirm BrowserAct tools are available to the intended Agent in this session before using them.  
- Saving the configuration or reaching the server URL alone does not confirm that the tools are ready.  
- If verification fails, explain what is missing and provide the next step.

### 4. Choose a first task

Once connected:

1. Show me my published Bots and the available official templates.  
2. If I have no published Bots, help me start with a suitable official template.  
3. Based on the Bots and templates actually available, suggest three tasks I can start with, plus an option to describe my own needs.

**Do not run any tasks yet.**

Before running a Bot or template, inspect its details and confirm the required inputs. When I ask you to run a task, check its progress and retrieve the results.

### Official client guide

Find and show the official MCP setup documentation for the AI client identified above. Use it for the client-specific setup steps. For general MCP client information:

[https://modelcontextprotocol.io/clients](https://modelcontextprotocol.io/clients)

### BrowserAct setup help

[https://docs.browseract.com/integrations/mcp](https://docs.browseract.com/integrations/mcp)

## From finding a Bot to getting results.

Access your published Bots, official templates, and task results through a single MCP server.

### Your published Bots

Find, inspect, and run your published Bots.

3 tools

- `search_bots`  
- `get_bot_details`  
- `run_bot`

### Official templates

Discover ready-made templates and run them with your inputs.

4 tools

- `search_official_templates`  
- `get_official_template_details`  
- `list_proxy_regions`  
- `run_official_template`

### Tasks & results

Check progress, retrieve results, or cancel a task.

4 tools

- `list_tasks`  
- `get_task_status`  
- `get_task`  
- `cancel_task`

## Only the tools you choose.

Create a custom MCP server that exposes selected Bot tools.

### Selected Bot tools

- Product prices  
- Lead research  
- Review collection

## More ways to connect.

Integrate BrowserAct with your favorite tools and automate your workflows.

### Frequently Asked Questions

**Do I need a BrowserAct API key?**

Yes. Add your API key to your client's MCP configuration using the Authorization: Bearer header. Keep your key in the configuration, rather than pasting it into a chat.

**Which AI clients can I use?**

Use a client that supports MCP. Choose your client in the setup section for its connection instructions. Configuration and remote-server support vary by client.

**What can the MCP server do?**

Find and run your published Bots, discover and run official templates, check task progress, retrieve results, and cancel tasks. The official server provides a fixed set of 11 tools.

**When should I use a custom server?**

Use a custom MCP server when you want to expose only selected Bot tools. Manage these selections in your BrowserAct account.

**Is MCP the same as Agent CLI?**

No. BrowserAct MCP connects AI clients to your cloud Bots and official templates. Agent CLI gives your local agent a browser it can control directly.

## Your next scraper starts here.

Describe the web data or task you want BrowserAct to build.

Find dental clinics in Austin on Google Maps. Return name, rating, review count, address, phone, website, and Maps URL.
