# ScreenshotOne MCP > ScreenshotOne MCP helps your AI agent capture, audit, and archive websites visually. Instead of opening a browser, your agent can instantly grab full-page screenshots or specific elements on any URL. It also generates professional PDFs from web pages, letting you check site designs, monitor content changes, or gather metadata without ever touching a manual workflow. ## Overview - **Category:** productivity - **Price:** Free - **Tags:** screenshot-api, pdf-generation, web-automation, visual-auditing, image-processing ## Description Managing visual content across multiple websites is usually a total pain. You have to open tabs, resize windows, and copy-paste screenshots just to verify if a button moved or if the layout changed. This MCP connects your agent directly to high-resolution web capture tools. When you use it through Vinkius, your agent handles the whole process; you simply ask for what you need in plain language. Need to audit how a site looks on mobile vs. desktop? Your agent can take screenshots of specific viewports or isolated elements. Want to archive an article? It grabs the full page and compiles it into a shareable PDF. Even if you just want to know the title and size of a page, your agent gets that data immediately without needing a full capture. This MCP turns complex visual auditing tasks—the kind of thing that used to take hours of manual clicking—into simple conversations with your AI client. ## Tools ### take_element_screenshot Takes an image of one specific part of the page, defined by its CSS selector. ### get_page_metadata Retrieves basic site information like titles and page size without taking any pictures. ### generate_pdf Creates a professional PDF document from a specified web address. ### take_full_page_screenshot Captures a single image that includes every pixel from the top edge to the bottom edge of the website. ### take_screenshot Takes a standard screenshot of an entire URL. ### take_viewport_screenshot Captures a screenshot sized to exact dimensions, simulating a specific device view. ## Prompt Examples **Prompt:** ``` Take a screenshot of https://vinkius.com. ``` **Response:** ``` I've captured the screenshot for you! You can access the high-quality image via the provided URL. Would you like a full-page capture or a PDF version as well? ``` **Prompt:** ``` Capture the full page of https://github.com/vinkius. ``` **Response:** ``` Full-page capture initiated! I've retrieved the image showing the entire GitHub profile. It includes all repositories and contributions. Would you like the direct link? ``` **Prompt:** ``` Show metadata for https://www.google.com. ``` **Response:** ``` I've retrieved the metadata for Google. The site title is 'Google' and the page size is approximately 50KB. Would you like me to take a screenshot now? ``` ## Capabilities ### Capture full-page site screenshots The agent takes an image showing the entire website, from top edge to bottom edge. ### Audit specific page elements You can capture just a small section of a webpage, like a contact form or a specific banner, without grabbing everything else. ### Create PDF reports from URLs The agent generates clean, professional PDFs that contain the full content of any given website address. ### Get site metadata instantly You can pull core data like the page title or overall file size without generating a single image or PDF. ### Capture specific viewports The agent takes screenshots scaled to precise dimensions, simulating different device screens. ## Use Cases ### The QA team needs to check for visual regressions. A developer pushes an update and the QA engineer asks their agent: 'Take a full-page screenshot of the checkout flow, then take three specific element screenshots (the payment button, the error message box) to verify the layout.' The agent uses `take_full_page_screenshot` followed by multiple calls to `take_element_screenshot`, giving instant visual proof. ### The marketing team needs to archive a press release. Instead of printing or taking messy screenshots, the archivist asks their agent: 'Generate a PDF from this URL.' The agent uses `generate_pdf` and provides a clean, professional document ready for immediate storage in a knowledge base. ### The developer needs to check site performance quickly. Before writing any code, the developer asks their agent: 'What is the page size of the homepage?' The agent uses `get_page_metadata`, returning the file metrics immediately so development can proceed without waiting for a full visual capture. ### The designer needs to see how content looks on different devices. The designer asks their agent: 'Show me what this page looks like in both 375px and 1440px viewports.' The agent uses `take_viewport_screenshot` multiple times, giving the designer precise visual data for design sign-off. ## Benefits - Verify design consistency instantly. Instead of manually resizing your browser window to check mobile vs. tablet views, use the `take_viewport_screenshot` tool to capture precise images for comparison. - Simplify content archiving. Generate professional PDFs from any URL using the `generate_pdf` tool, giving you a reliable, shareable record without needing manual screen grabs. - Focus on what matters with targeted shots. When auditing specific components, use `take_element_screenshot` to isolate and capture only the necessary CSS element. - Get data fast, no pictures needed. Use `get_page_metadata` to pull titles, descriptions, and page sizes in a single API call, skipping the visual step entirely. - Capture the full context every time. The `take_full_page_screenshot` tool ensures you get a complete view of long-form content or complex site layouts, eliminating partial captures. ## How It Works The bottom line is that you talk to your AI client, and it does the visual work for you. 1. Connect your AI client to the ScreenshotOne MCP using your access key. 2. Tell your agent exactly what you need—for example, 'Give me a PDF of this URL' or 'Take a full-page screenshot of X'. 3. Your agent uses the tools to perform the capture and returns the high-quality image link or generated file. ## Frequently Asked Questions **How does ScreenshotOne MCP handle full-page captures?** The agent uses the `take_full_page_screenshot` tool to generate one single image that stretches from the very top of the page content to the bottom. You get a complete visual context, which is great for archiving entire articles. **Can ScreenshotOne MCP capture just a specific section?** Yes. If you only need the header or a contact form, use the `take_element_screenshot` tool. You specify the exact CSS element, and the agent captures nothing else, keeping your image clean. **Is ScreenshotOne MCP better than just taking a regular picture?** Absolutely. This MCP provides programmatic access to high-resolution web data, allowing you to capture specific viewports or metadata programmatically. It's not limited by your screen size. **How do I get a PDF using ScreenshotOne MCP?** You simply instruct the agent to 'Generate a PDF from this URL.' The `generate_pdf` tool handles the conversion, giving you a clean, structured document file ready for sharing. **Does ScreenshotOne MCP work if I don't know what page size I need?** You can use `get_page_metadata` first to understand the site's dimensions. Then, you can ask the agent to take a viewport screenshot using those retrieved metrics. **How do I find my ScreenshotOne Access Key?** Log in to your [**ScreenshotOne dashboard**](https://screenshotone.com/app/settings), and you will find your Access Key under the 'API Keys' section. Copy and paste it below. **Can the agent capture elements by ID or Class?** Yes. Use the `take_element_screenshot` tool providing a valid CSS selector (e.g., `#header` or `.product-card`). Your agent will return a screenshot of just that element. **Is PDF generation supported for all sites?** Yes. The `generate_pdf` tool works for any publicly accessible URL, converting the website content into a high-quality PDF document instantly.