# Fluxguard MCP MCP > Fluxguard monitors your live websites for unexpected changes. It tracks text, HTML structure, and visual appearance across monitored pages. When a site breaks or looks different from the last snapshot, Fluxguard alerts your AI agent immediately. You can manage all monitoring setups—adding URLs, grouping them into categories, or manually crawling sites—all without leaving your chat window. ## Overview - **Category:** security-compliance - **Price:** Free - **Tags:** website-monitoring, regression-testing, change-detection, visual-testing, alerts ## Description When you connect this MCP to Vinkius, your agent gets a dedicated connection to Fluxguard's entire web monitoring platform. Instead of checking dashboards and dealing with spreadsheets, you just ask your AI client what changed on the site. It pulls in all detected variations—whether it’s a headline that got reworded or an image that shifted its position. You can even trigger immediate crawls for specific pages if you suspect something is wrong. This process lets you manage entire sets of URLs and categorize them by function, keeping your monitoring setup clean. You retrieve alerts as they happen, get detailed breakdowns of every detected change, and review historical snapshots to prove when the break occurred. ## Tools ### acknowledge_alert Marks a specific monitoring alert as reviewed and resolved in the system. ### add_page Adds a new URL to the monitored list so changes can be tracked over time. ### create_category Establishes a logical group name for related URLs within your monitoring setup. ### get_account Retrieves general details about the connected Fluxguard account and organization settings. ### get_change Pulls detailed information about a specific detected change event. ### get_site Retrieves general details and status for a monitored site. ### initiate_crawl Triggers an immediate, manual scan of the specified website URL. ### list_alerts Fetches a comprehensive list of all active and past monitoring alerts. ### list_categories Returns an organized list of the custom categories you've created for monitoring pages. ### list_changes Lists all detected changes across monitored sites, giving a summary view. ### list_sites Retrieves the list of all URLs and sites currently under monitoring. ### list_snapshots Shows a history of captured site snapshots, allowing you to review past versions. ## Prompt Examples **Prompt:** ``` Add the URL 'https://example.com' to my monitoring list. ``` **Response:** ``` Adding page... 'https://example.com' has been successfully added to your monitoring list. ``` **Prompt:** ``` Show me the most recent changes detected across all sites. ``` **Response:** ``` Retrieving changes... I found 3 recent changes across your monitored pages. ``` **Prompt:** ``` Initiate an immediate crawl for site ID 'site_123'. ``` **Response:** ``` Crawl initiated! I've started an immediate crawl for site ID 'site_123'. ``` ## Capabilities ### Add Pages for Monitoring Instructs the agent to add a new URL that needs continuous monitoring. ### Create Organization Categories Groups multiple monitored URLs into logical categories (e.g., 'Checkout Flow' or 'Marketing Pages'). ### Manually Trigger Site Crawls Forces an immediate, on-demand crawl of a specific site to check for recent changes. ### List and View Alerts Retrieves the full list of all detected website changes and lets you mark them as reviewed. ### Retrieve Change Details Fetches specific, granular data about a particular change event or alert. ### Manage Historical Snapshots Lists and retrieves past captured versions of monitored pages for comparison. ## Use Cases ### The Marketing Page Break A marketing manager sees an ad pointing to the 'Pricing' page, but the main headline is wrong. The agent runs `initiate_crawl` on the URL and then uses `list_changes`, showing that the H1 tag text changed from 'Plans for Everyone' to 'Basic Tiers Only.' Problem solved in two steps. ### Post-Deployment Audit A developer just deployed a new feature and needs QA sign-off. They run `list_sites` first, then ask the agent to review all alerts for the 'User Profile' section using `get_change`, ensuring no unintended visual regressions were introduced. ### Investigating Old Bugs Support needs to know if a layout issue was fixed last month. They ask the agent to list historical snapshots (`list_snapshots`) for that specific page, letting them visually compare the current site against the snapshot taken on the date of the original bug report. ### Cleanup After Incident The SRE team found 50 alerts overnight. They run `list_alerts` to get the full count, then use `acknowledge_alert` repeatedly as they investigate each one, keeping a clear record of what's been reviewed. ## Benefits - Immediate Visibility: Instead of waiting for a scheduled check, you can manually trigger crawls using `initiate_crawl` to test specific pages right now. This is crucial when investigating an urgent production bug. - Structured Alert Management: The agent lets you review all potential issues via `list_alerts`, and then confirms resolution by calling `acknowledge_alert`. It keeps your alert history clean and actionable. - Deep Change Analysis: If a change pops up, don't just see 'it changed.' Use `get_change` to pull the full details—HTML differences, visual metrics, everything you need to debug it fast. - Organization Control: Group related pages into logical buckets using `create_category`. This helps your agent filter results when you only care about, say, 'Payment Flow' changes, ignoring everything else. - Historical Proof: You can prove *when* something went wrong. Listing and retrieving snapshots with `list_snapshots` gives you a clear, timestamped record of the site's previous state. ## How It Works The bottom line is you talk to your site monitoring tools like you're talking to a teammate, skipping the web UI entirely. 1. First, connect your Fluxguard API key to the MCP via Vinkius. This gives your AI agent access to all monitoring tools. 2. Next, tell your agent what you need: 'Add this URL' or 'List all alerts.' The agent executes the request and pulls data from Fluxguard. 3. Finally, your agent presents the information—a list of detected changes, a summary of an alert, or a historical snapshot—directly into the chat for review. ## Frequently Asked Questions **How do I find out what changed on a page using Fluxguard MCP?** The agent uses the `list_changes` tool to pull all detected shifts. You can then use `get_change` if you need deep details about one specific change event. **Can I make my monitoring setup more organized using Fluxguard MCP?** Yes, use the `create_category` tool to group related URLs. This helps your agent filter results so you only see changes relevant to a specific section of your site. **What if I need to check a page right now and don't want to wait for the next scheduled crawl? (initiate_crawl)** You can force an immediate check by calling `initiate_crawl`. This manually triggers a scan of your site, giving you current data instantly. **How do I mark a detected issue as resolved after investigating it?** After confirming the issue is fixed or reviewed, run the `acknowledge_alert` tool. This marks the alert in Fluxguard's system so you don't have to look at it again. **Does Fluxguard MCP track how my site looked last week? (list_snapshots)** Yes, using `list_snapshots` allows you to see a history of captured versions. This is essential for proving when a visual regression occurred relative to previous builds. **What information can I retrieve using the `get_account` function?** It pulls your organization's high-level attributes. This is useful for verifying API key status or checking overall account limits before you start monitoring pages. **When I use `get_change`, what granular details does it provide about a detected modification?** It gives deep data on the change, not just a summary. You'll get specific text differences, HTML snippets, and timestamps for pinpoint accuracy. **What is the proper way to use `add_page` when setting up a new monitored website?** You simply provide the full URL you want to track. The system adds it to your monitoring queue and lets you assign it to a category right away. **How do I get an API Key for Fluxguard?** You can generate an API key in your Fluxguard dashboard under Organization Settings. **What types of changes can Fluxguard detect?** Fluxguard detects text, HTML, and visual changes (CSS/layout) across monitored pages. **Can I manually trigger a check for a site?** Yes, you can use the 'initiate_crawl' tool to trigger an immediate crawl for any monitored site. **Can I acknowledge alerts through the agent?** Yes, use the 'acknowledge_alert' tool to mark an alert as reviewed, removing it from the active list.