# Browse AI MCP for AI Agents MCP

> Browse AI connects web scraping, data extraction, and website monitoring into a single conversational tool for your AI agents. Instead of building complex pipelines, you can ask your agent to track competitor prices, gather product details from multiple pages, or monitor changes on key industry sites—all through natural language conversation.

## Overview
- **Category:** data-management
- **Price:** Free
- **Tags:** data-extraction, web-monitoring, no-code, robot-automation, data-retrieval, scraping-tasks

## Description

Need reliable data pulled off the web without writing code? This MCP lets you talk to your agent and get structured information straight into your workflow. You can tell it to check a website for specific price points, run automated reports across multiple sources, or list every robot scraper you own. It handles the complicated part of managing scraping jobs—from triggering runs on URLs to pulling back the final data.

If you're used to building complex scripts just to monitor a few prices, this changes that. You simply connect your agent via Vinkius and ask for what you need. Your agent then uses these tools to manage monitors, run tasks against specific websites, and retrieve structured data into a format ready for analysis. It’s all about automating the tedious work of web data capture.

## Tools

### create_monitor
Sets up a scheduled task that automatically tracks changes on a designated website for a specific robot.

### get_bulk_run
Retrieves detailed information about one specific bulk data extraction job.

### get_robot
Fetches the full metadata and details for a single, named web scraping robot.

### get_system_status
Checks the overall operational status of the browsing AI infrastructure and its current task queue.

### get_task
Retrieves the final data and status report for a single, ongoing web scraping task.

### list_bulk_runs
Provides an overview of all previously executed bulk data extraction jobs across various sources.

### list_monitors
Shows a list of all existing, scheduled website monitors tied to a specific robot.

### list_robots
Retrieves the names and metadata for every approved web scraping robot you have set up.

### list_tasks
Lists all tasks associated with a specific robot, helping you track multiple running jobs.

### run_robot
Starts a data extraction process by running a specified web scraping robot against a target URL, creating a new task.

## Prompt Examples

**Prompt:** 
```
List all my approved web scraping robots.
```

**Response:** 
```
**✅ Your Robots:**
*   Amazon Price Tracker (ID: rob_1)
*   LinkedIn Lead Extractor (ID: rob_2)
*   Real Estate Monitor (ID: rob_3)

These scrapers are active and ready to pull data. Which one should we run first?
```

**Prompt:** 
```
Run robot 'rob_2' on the new lead page and get me the data.
```

**Response:** 
```
**🚀 Task Initiated:**
Robot `rob_2` is now processing the URL. We created task ID `task_998877`. 

*Status:* Processing...
*ETA:* 5 minutes.

You can check progress or pull results using this ID.
```

**Prompt:** 
```
What data did task task_99283 pull?
```

**Response:** 
```
**✅ Task Status: SUCCESSFUL**
Task `task_99283` is complete. Here’s the structured summary:

| Field | Value |
| :--- | :--- |
| Product Name | Modern Lamp |
| Price | $45.00 |
| Stock | In Stock |

Would you like me to generate a JSON file with all this data?
```

## Capabilities

### Manage Web Scrapers
Get a list of your existing scraping robots and view detailed information about each one.

### Execute Data Capture Tasks
Initiate data extraction runs on specific web pages, then monitor the task status until the structured results are ready for you to pull.

### Monitor Websites Automatically
Set up scheduled alerts that track changes on any given website, so you don't have to check it manually every day.

### Review Bulk Data Jobs
View and manage runs where data was extracted from multiple sources simultaneously, giving you a high-level overview of all the captured information.

### Check System Health
Verify the operational status of the web scraping infrastructure and check current task queues.

## Use Cases

### Tracking Competitor Pricing Changes
A growth marketer asks their agent, 'Set up monitoring for competitor prices.' The agent uses the `create_monitor` tool to track specific URLs, alerting the user immediately if a price changes by more than 10%.

### Gathering Lead Information from Job Boards
A data analyst tells their agent to 'Run the lead extractor robot on today's top job board.' The agent uses `run_robot` and then retrieves the captured structured data via `get_task`, giving a clean list of contact details.

### Auditing Multiple Website Data Sources
A researcher needs to compare product specs across five different vendor sites. They ask their agent to manage and inspect the results using `list_bulk_runs` after running several tasks, comparing all data points at once.

### Verifying Scraping Capability
A developer needs to know if their scraping setup is working. They ask the agent for a system check, and it uses `get_system_status` to confirm that the whole web data infrastructure is online.

## Benefits

- Stop exporting dashboards manually. Use 'get_task' to pull structured data directly into your conversation, eliminating manual copy-pasting.
- Monitor competitors effortlessly. Set up a monitor schedule using 'create_monitor' so your agent alerts you immediately when prices or listings change.
- Manage complex operations easily. Run multiple sources at once and view the results by calling 'list_bulk_runs', keeping all your data in one place.
- Control every aspect of scraping. From listing available scrapers with 'list_robots' to initiating a run via 'run_robot,' you control the whole process conversationally.
- Check infrastructure health instantly. Use 'get_system_status' when you need to know if the web data tools are running smoothly or if there’s a queue backup.

## How It Works

The bottom line is: You use natural language conversation to control powerful, underlying web automation processes.

1. Subscribe to this MCP and supply your unique Browse AI Secret API Key.
2. Connect your agent via any MCP-compatible client, granting it access to the web data tools.
3. Ask your agent to perform an action, like 'Run a monitor on the competitor's pricing page,' and retrieve the results.

## Frequently Asked Questions

**How can I automatically track changes on competitor websites using Browse AI?**
You can set up scheduled monitoring for specific pages. The system tracks defined elements (like pricing or listings) and alerts you instantly if the content deviates from what was previously captured. This saves you constant manual checks.

**Do I need to write code to extract data with Browse AI?**
No, you don't. You just talk to your agent and tell it which data points you want from a given URL or set of websites. The MCP handles the complex scraping logic in the background.

**What is the difference between running a task and bulk runs with Browse AI?**
A single 'task' extracts data from one specific URL at a time. Bulk runs, however, are designed to pull data across many different sources or pages simultaneously, giving you a comprehensive report.

**Can I use Browse AI MCP to gather leads from job boards?**
Yes. You can instruct your agent to run specific scraping robots on large websites like job boards. The MCP extracts the required data, such as names and contact info, into a usable list.

**Does Browse AI MCP work with my existing code?**
Absolutely. You can integrate web extraction logic directly into your coding environment by having your agent manage the tasks and retrieve results through conversational commands.