# Swiftype MCP for AI Agents AI Agent Connect > Swiftype. This Connector lets your AI client talk directly to your Swiftype and Elastic search engines. You can query live data, manage your document index, and pull real-time analytics like top clicks without leaving your prompt. It turns your AI into a hands-on data architect for your search infrastructure. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_G3YqlNBTOlJoB2C6UQsglvpbCGCydRyab9CjoxGh/ai-agent-connect - **Tags:** enterprise-search, site-search, indexing, search-api, data-retrieval ## Description Swiftype is an Connector that lets you talk directly to your Swiftype and Elastic search engines. Most people find managing a search index a headache because it usually means jumping between different web dashboards and manually tweaking JSON records. This connector changes that by letting your AI do the heavy lifting for you. Instead of copy-pasting IDs or hunting for specific search engine metrics, you just tell your agent what you need. You can have it build out new document structures, wipe out old ones, or dig into how people are actually using your search results. It's a huge time-saver for anyone who needs to keep their search data clean and accurate without the constant tab-switching. You can find this in the Vinkius catalog to get it hooked up quickly. It makes search administration feel like a conversation rather than a chore. It handles the tedious parts of data management like checking crawler limits or auditing your schemas. When you can ask your agent to pull top click data or search trends on the fly, you stop guessing what your users want and start seeing it in real time. This means your content stays relevant and your search results stay sharp without you having to do the manual digging every single day. ## Tools ### st.create_documents Create new structured records in your search index. This lets you push new data into your engine without manual entry. ### st.delete_documents Permanently remove specific documents or cached pages from your search results. Use this to clean up old records or fix errors in your index. ### st.list_documents View all stored metadata and IDs for a specific document type. This helps you audit your records and see exactly what is indexed. ### st.list_domains Check crawler limits and identify active arrays in your index scopes. You can use this to verify that your automated crawlers are staying within bounds. ### st.list_engines See a list of all active isolated Elastic indices bound to your account. This gives you a clear view of your entire search infrastructure. ### st.post_search Fire raw queries into a specific engine to get back faceted JSON results. It allows your agent to pull precise data based on specific user searches. ### st.post_suggest Get predictive autocompletes and spelling-tolerant matches for your search bar. This helps users find what they need faster by offering smart suggestions. ### st.analytics_top_clicks Find out which documents are receiving the most clicks from users. Use this to identify your highest-performing content. ### st.analytics_top_searches Look at the most frequent search queries being used by your customers. This provides a direct look at what your audience is trying to find. ### st.list_doc_types Pull the schema blueprints that define your data types and vault limits. Use this to understand the structure of your data and its constraints. ## Prompt Examples **Prompt:** ``` Can you show me which search engines I have active and what kind of documents they hold? ``` **Response:** ``` I've checked your Swiftype account. You have 3 active engines: - **Global-Help-Center**: Holds documentation and FAQ records. - **Ecommerce-Catalog-US**: Holds product listings and pricing. - **Internal-Company-Wiki**: Holds internal HR and policy docs. Would you like me to pull the specific schema for any of these? ``` **Prompt:** ``` Run a search for 'refund policy' in the help center and tell me the top 3 most clicked results. ``` **Response:** ``` I'm searching the 'Global-Help-Center' engine for 'refund policy'... Found 12 relevant documents. Here are the top 3 most clicked results: | Document | Clicks | | :--- | :--- | | /help/refund-process | 142 | | /docs/refund-eligibility | 88 | | /support/refund-status | 54 | ``` **Prompt:** ``` Check the crawler limits for my active domains. ``` **Response:** ``` I've checked the crawler limits across your active index scopes. - **Domain A**: 500 requests/min - **Domain B**: 1000 requests/min Your current crawler is operating within these bounds. Let me know if you need to adjust any specific array limits. ``` ## Capabilities ### Query live search results Get faceted JSON data from specific engines instantly. ### Bulk create documents Inject new data payloads into your index in one go. ### Delete old records Remove specific documents or cached pages from your search results. ### Fetch search analytics See what queries are actually being typed by your users. ### Track top clicks Identify which documents are getting the most engagement. ### Inspect engine schemas Browse your active search engines and their data types. ## Use Cases ### Analyzing content performance A marketing manager needs to know why a specific help article isn't getting traffic. They ask the agent to run st.analytics_top_searches and st.analytics_top_clicks to see if people are even finding the right terms. ### Purging test data An engineer needs to clear out old test data from a production index. They tell the agent to use st.delete_documents to purge specific IDs without touching the rest of the map. ### Bulk indexing new FAQs A content team just finished a big batch of new FAQs. They have the agent use st.create_documents to inject all the new JSON payloads into the 'help-center' engine at once. ### Checking crawler limits A dev wants to see if the crawler is hitting limits. They ask the agent to use st.list_domains to check the active arrays and crawler limits across all scopes. ## Benefits - Skip the dashboard: Use st.list_engines and st.list_doc_types to explore your entire search architecture without logging into a web portal. - Fast data updates: Use st.create_documents to push new content into your index in bulk rather than one by one. - Precise data cleaning: Use st.delete_documents to quickly vaporize old records or clear out cached pages that are messing up your results. - Real-time user insights: Pull live data using st.analytics_top_searches to see exactly what your customers are looking for right now. - Conversion tracking: Use st.analytics_top_clicks to identify high-performing content and see which search paths are actually working. - Predictive search: Implement smart autocompletes by using st.post_suggest to help your users find what they need faster. ## How It Works The bottom line is that you get a direct line of communication between your AI and your search data. 1. Add the Swiftype MCP to your Vinkius integration settings. 2. Provide your SWIFTYPE_API_KEY from your Elastic Swiftype profile. 3. Ask your agent to list engines, run searches, or pull analytics. ## Frequently Asked Questions **What can I do with the Swiftype MCP?** I can use it to manage your Swiftype search index directly. This means your AI agent can create new documents, delete old ones, run searches, and pull real-time analytics like top clicks and searches without you having to log into a dashboard. **Does Swiftype MCP work with Elastic App Search?** Yes, it's designed to connect your AI to Swiftype and Elastic App Search. It lets your agent interact with your search engines to perform tasks like schema discovery and bulk data indexing. **Can I use this to see what my customers are searching for?** Absolutely. You can ask your agent to pull the top search queries and click data. It helps you see what's trending and which content is actually converting your users. **How do I manage my search index data with this?** You can use the Connector to perform CRUD operations. Your agent can inject new JSON payloads into your index, list all your current documents, or vaporize specific records that are no longer needed. **Is this Connector good for content marketing?** It's great for it. Instead of guessing what users want, you can use it to pull real-world click data and search trends to see which topics are actually resonating with your audience. **Can it help me fix search errors?** Yes, it helps with data reliability. You can use it to audit your schema blueprints and inspect your active engines to find and resolve mapping constraints or index bugs. **Can it delete documents from my search engines?** Yes. The `st.delete_documents` tool removes specific documents by ID from an engine. Each deletion requires explicit document IDs, so nothing is removed accidentally. **Does it support search analytics like CTR?** Yes. Use `st.analytics_top_searches` and `st.analytics_top_clicks` to retrieve real usage data — top queries, click-through rates, and conversion metrics. **What Swiftype plan is required?** Any plan that includes Site Search API access. You need a valid API key from your Swiftype dashboard — the integration uses standard REST endpoints available to all API-enabled accounts.