# Nimbleway MCP

> Nimbleway lets your AI client scrape and search the live web right from your chat window or IDE. Extract content—whether you need raw HTML or clean Markdown—and perform structured searches against major search engines. It's a direct connection to high-performance data collection, letting your agent fetch information without needing external tools or browser extensions.

## Overview
- **Category:** developer-tools
- **Price:** Free
- **Tags:** data-extraction, proxy-services, web-crawling, stealth-browsing, structured-data, automated-search

## Description

Nimbleway connects your AI client directly to web scraping and advanced search capabilities. You can ask your agent to fetch any webpage, rendering it as either raw HTML or clean Markdown format while bypassing typical anti-bot protections. Need market intelligence? Use the MCP to perform structured searches across major search engines and get organized data back immediately. Beyond simple extraction, you manage complex operations: check account usage via `get_account_usage`, monitor active jobs using `list_jobs`, and even inspect proxy configuration details with `list_proxies`. If your AI client is compatible with Vinkius, connecting this MCP gives you access to all these web data tools through a natural conversation. This means complex tasks—like checking pipeline status via `get_pipeline` or viewing account info with `get_me`—become simple commands for your agent.

## Tools

### extract_html
Pulls the full raw HTML content from a specified webpage URL.

### extract_markdown
Extracts the clean, readable text and formatting (like headers) from a webpage as Markdown.

### get_job
Retrieves specific status details for an individual scraping job run.

### get_me
Pulls general current account information and metadata about your Nimbleway profile.

### get_pipeline
Provides detailed status and configuration information for a specific data scraping pipeline.

### get_account_usage
Checks your account's current bandwidth usage, remaining credits, and overall billing status.

### list_jobs
Generates a list of all completed and active scraping jobs you have run.

### list_pipelines
Lists every available data collection pipeline you've set up in your account.

### list_proxies
Accesses the configuration details for all residential and data center proxies associated with your account.

### search_web
Executes a structured search query against major web engines and returns organized results.

## Prompt Examples

**Prompt:** 
```
Extract the content of 'https://example.com' as Markdown.
```

**Response:** 
```
Initializing extraction... I've successfully fetched the page and converted it to Markdown. The content includes headers, lists, and links from the original site. Shall I summarize it for you?
```

**Prompt:** 
```
Search the web for 'latest AI developments 2024' and give me structured results.
```

**Response:** 
```
Executing web search... I've found several top results regarding generative AI, LLM improvements, and new hardware releases. I've retrieved the snippets, titles, and URLs for each. Would you like to extract the full content of the top result?
```

**Prompt:** 
```
Check my account usage and remaining credits.
```

**Response:** 
```
Retrieving usage data... You have used 15GB of bandwidth this month out of your 50GB limit. You also have 4,500 successful extraction credits remaining. Your account status is 'ACTIVE'.
```

## Capabilities

### Scrape Web Content
Request the content of any live webpage, receiving it either as raw HTML or clean Markdown text.

### Perform Structured Searches
Execute real-time web searches and receive structured data points from major search engines.

### Monitor Data Jobs
List, check the status of, or retrieve metadata for active scraping jobs.

### Manage Scraping Pipelines
View and inspect details about your configured data extraction workflows.

### Check Account Status
Retrieve current bandwidth usage, remaining credits, and overall account metrics.

## Use Cases

### Competitive Analysis
A market researcher needs competitor pricing and product features across five different sites. Instead of opening five browser tabs, they ask their agent to use `extract_markdown` on all five pages and then compile the data into a single comparison table.

### Building Agent Proofs
A developer is building an automated content generator. They first use `search_web` for current industry trends, then use `extract_html` on the top result to gather raw source material, all within a single session.

### Auditing Web Crawl Jobs
A data engineer suspects a pipeline is failing. They immediately ask the agent to run `list_jobs`, find the ID, and then use `get_job` to diagnose exactly why the last extraction failed.

### Real-Time Information Gathering
A student needs quick facts about a niche topic. Instead of doing multiple Google searches, they ask their agent to run `search_web`, which returns structured results that are immediately usable in an essay draft.

## Benefits

- Go beyond simple data copying. Use `search_web` to get organized, structured results from live web searches, which is critical for market research.
- Avoid messy code blocks. Instead of writing complex scraping scripts, simply ask your agent to use `extract_markdown` and receive clean, readable text directly in the chat.
- Know your limits before you start. Use `get_account_usage` to check bandwidth and remaining credits instantly so you don't hit a paywall mid-project.
- Manage large operations without leaving your IDE. You can use `list_pipelines` or `list_jobs` to monitor complex data streams right alongside your code development.
- Bypass common anti-bot measures. The underlying technology handles the tricky rendering, allowing you to focus purely on the data you need from a given URL.

## How It Works

The bottom line is that instead of writing code to connect to scraping APIs, you just tell your AI client what web data you need.

1. Subscribe to this MCP and provide your specific Nimbleway API Key in your client settings.
2. Your AI agent recognizes the available web data tools (like `extract_markdown` or `search_web`) within Vinkius.
3. You issue a natural language command, and your agent invokes the tool, executing the scrape or search and returning the structured data to you.

## Frequently Asked Questions

**How does Nimbleway MCP handle website anti-bot measures?**
Nimbleway uses advanced stealth technology to bypass common bot detection systems. This lets your agent scrape protected sites and retrieve content you otherwise couldn't access.

**Can I use Nimbleway MCP for general web searching?**
Yes, the `search_web` tool performs structured searches across major search engines, giving you organized snippets, titles, and URLs immediately.

**What is the difference between extract_html and extract_markdown?**
Use `extract_markdown` when you want clean, human-readable text (like an article). Use `extract_html` if you need the raw source code structure for development purposes.

**How do I check my usage with Nimbleway MCP?**
You call the `get_account_usage` tool. This instantly reports your current bandwidth used, remaining credits, and overall account status so you know exactly how much data you can pull.

**Is Nimbleway MCP just for scraping one site?**
No, it's designed for variety. You can use `search_web` for broad research, and then use tools like `extract_markdown` on the specific sites that interest you.