# Constructor MCP for AI Agents AI Agent Connect > Constructor MCP lets you manage your e-commerce search and product discovery workflows directly from your AI agent. It handles ML-ranked product searches, personalized recommendations, and category browsing. Use it to audit your site's search logic, test attribute filters, or check marketing collections without leaving your workspace. It connects your Constructor.io account to your agent for instant feedback on discovery performance. ## Overview - **Category:** ecommerce - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_7QdndEEXMElkrXV72k1k9X9VY0HS8rfUr6VIhrH5/ai-agent-connect - **Tags:** site-search, product-discovery, personalization, recommendation-engine, nlp, e-commerce-optimization ## Description Connecting your Constructor account to your AI agent changes how you manage your online store. Instead of jumping between different tabs to check how a search query performs or how a specific recommendation pod is behaving, you can just ask your agent to do the heavy lifting. You can see how products rank in real-time, check if your brand taxonomies are actually showing up for customers, and verify that your size and color filters are working as intended. It turns a manual auditing process into a conversation. Whether you're debugging a specific search parameter or checking a curated collection for a new marketing campaign, this Connector gives you a direct line into your site's discovery engine. It's a huge time saver for anyone who needs to keep their e-commerce experience sharp. You can find this and thousands of other tools in the Vinkius catalog to build out your specific tech stack. This setup means you spend less time in dashboards and more time actually improving how people find what they want to buy. It's a direct way to get answers about your store's logic without the friction of a traditional UI. You can quickly audit your entire site search to ensure that your most profitable items are appearing where they should. It removes the need to manually type out dozens of queries just to see if your ML ranking is behaving correctly. You get a clear, text-based summary of your catalog's behavior, making it much easier to spot errors in your category mapping or recommendation logic before your customers do. ## Tools ### autocomplete Get the logic behind your active autocomplete properties. This helps you see how your search bar suggests items to users in real time. ### browse_collection Identify active arrays for native gateway auth. Use this to find specific marketing clusters or curated groups. ### custom_search Find active arrays spanning specific hold parsing. This helps you identify precise search results for complex or custom queries. ### search_filtered Apply JSON structures to restrict results by color, size, or features. Use this to verify that your attribute filters are working correctly for shoppers. ### search_pagination Run automated validation checks on gateway history. This lets you check how your search results are paginated across different pages. ### get_recommendations Retrieve cloud logging for vault limits. Use this to see what items are appearing in your recommendation pods and personalized feeds. ### search_products Find CRM records inside the headless platform. This lets you see the actual product results and rankings for any given search term. ### search_sorted List attached rules for exporting active billing. Use this to check the sorting rules applied to your product lists and search results. ### browse_brand Inspect internal arrays for specific brand data. Use this to check how your brand taxonomies are structured for your customers. ### browse_category Fetch a JSON payload for specific category bindings. This is useful for auditing your product directory trees and navigation. ## Prompt Examples **Prompt:** ``` Show me the top 5 products for 'waterproof jackets'. ``` **Response:** ``` ### Search Results for 'waterproof jackets' I found the top 5 ML-ranked items for your query: | Product Name | Price | Status | | :--- | :--- | :--- | | **Apex Waterproof Shell** | $189.00 | In Stock | | **Trail Blazer Jacket** | $155.00 | In Stock | | **City Rain Guard** | $120.00 | Low Stock | | **Summit Shield** | $210.00 | In Stock | | **Eco-Dry Parka** | $195.00 | In Stock | ``` **Prompt:** ``` What's in the 'new-arrivals' collection? ``` **Response:** ``` ### Collection: New Arrivals Here are the items currently mapped to the **New Arrivals** collection: * **Luna Running Shoes** (ID: prod_992) * **Solaris Backpack** (ID: prod_881) * **Mist Yoga Mat** (ID: prod_772) * **Zenith Water Bottle** (ID: prod_663) *Total items found: 4* ``` **Prompt:** ``` Are the 'Blue' and 'XL' filters working for running shoes? ``` **Response:** ``` ### Filter Verification: Running Shoes I applied the filters for **Color: Blue** and **Size: XL**. Here are the results: * **Blue Speedster (XL)** - Found * **Midnight Blue Runner (XL)** - Found * **Ocean Breeze Trainer (XL)** - Found **Status:** ✅ The attribute filters are correctly mapping to the product catalog for this category. ``` ## Capabilities ### Run ML-ranked product searches Ask your agent to pull the top products for any search term to see how your ranking logic performs. ### Get predictive autocomplete data Check the logic behind your search bar's suggestions to ensure they match user intent. ### Fetch personalized recommendations See what items are appearing in specific recommendation pods and personalized feeds. ### Browse brand and category hierarchies Inspect your product directory trees and brand taxonomies to ensure correct navigation mapping. ### Apply complex attribute filters Verify that shoppers can correctly filter by size, color, and other specific features. ### Audit curated marketing collections Identify and review the contents of your marketing clusters and static collections. ## Use Cases ### Checking search rankings for new products A manager wants to see if 'hiking boots' shows the correct items. The agent uses search_products to pull the top 5 results and check the ranking. ### Verifying marketing collection content A marketing coordinator needs to check if the 'Summer Sale' items are live. The agent uses browse_collection to list the items in that group. ### Debugging attribute filters A developer needs to know why 'Blue' isn't showing up in shoes. The agent uses search_filtered to test the specific attribute logic. ### Auditing recommendation pods A product owner wants to see what's in the 'Trending' section. The agent uses get_recommendations to pull the current personalized items. ## Benefits - Stop manual dashboard testing by using search_products to check rankings and results instantly. - Verify your marketing strategy with browse_collection to see if your curated clusters are live. - Fix filtering bugs faster by using search_filtered to test specific attributes like size or color. - Improve user experience by checking autocomplete logic to ensure your search bar suggests the right items. - Audit your site's hierarchy using browse_category and browse_brand to ensure correct taxonomy mapping. ## How It Works The bottom line is you get a direct text interface for your entire Constructor.io discovery engine. 1. Subscribe to the Constructor MCP via the Vinkius catalog. 2. Provide your Constructor Public API Key from your dashboard settings. 3. Ask your agent to check search results, filter logic, or recommendation data. ## Frequently Asked Questions **Can I use the Constructor MCP to check my search rankings?** Yes, you can ask your agent to pull the top results for any query to see how your ML ranking is performing in real-time. **Does the Constructor MCP help with personalized recommendations?** It lets you see what products are being served in specific recommendation pods and personalized feeds without leaving your chat. **How do I check my brand taxonomies?** You can ask the agent to browse your brands to see how they are structured in the backend, ensuring your navigation is correct. **Can I use this to test my site filters?** Yes, you can use it to verify that specific attributes like size or color are returning the correct products for your shoppers. **Is the Constructor MCP good for marketing audits?** It's great for checking that your curated collections and marketing clusters are correctly mapped and visible to customers. **Can my agent check the ML ranking for a specific product search?** Yes. Use the 'search_products' tool. The agent will retrieve results ranked by Constructor's ML engine, allowing you to audit how products are surfaced based on specific keywords and intent signals. **How do I retrieve personalized recommendations via the agent?** Provide the 'pod_id' to your agent and use the 'get_recommendations' tool. The agent will query the collaborative filtering models to return a list of products tailored to your specified recommendation logic. **Can I test attribute filtering like color or size through chat?** Absolutely. The 'search_filtered' tool allows you to pass exact attribute mappings (e.g., 'color:blue,size:L'). Your agent will verify how the API restricts results to those specific structural bounds.