Swiftype MCP. Run deep queries and manage indexes directly from your agent.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Swiftype connects your AI agent directly to Elastic Search instances. It lets you manage document indexes, run complex queries against custom content engines, and pull deep analytics data like top clicks—all without opening a web browser or writing manual API calls.
What your AI agents can do
St.analytics top clicks
Identifies active arrays detailing which documents received the highest number of clicks.
St.analytics top searches
Inspects internal data to show deep arrays listing the most frequent search terms used by users.
St.create documents
Adds new structured documents to your index using a list of explicitly attached rules.
You can run st.list_engines to get a list of every isolated Elastic index bound to your account.
Use st.post_search to fire raw queries into specific engines, or use st.post_suggest for predictive autocompletion on a given term.
The agent can list all stored metadata (st.list_documents), create new documents in bulk (st.create_documents), or permanently remove them (st.delete_documents).
Pull real-world data using st.analytics_top_searches and calculate high-conversion clicks with st.analytics_top_clicks.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Swiftype: 10 Tools for Deep Data Retrieval
These ten tools let your AI agent perform every administrative task required to interact with an enterprise search stack—from simple suggestions to complex CRUD operations.
019d760fst.analytics top clicks
Identifies active arrays detailing which documents received the highest number of clicks.
019d760fst.analytics top searches
Inspects internal data to show deep arrays listing the most frequent search terms used by users.
019d760fst.create documents
Adds new structured documents to your index using a list of explicitly attached rules.
019d760fst.delete documents
Permanently removes cached pages and data metrics from search indices after validation.
019d760fst.list doc types
Extracts schema blueprints, showing you what types of data are stored in your indexes.
019d760fst.list documents
Dumps all stored metadata, allowing you to track IDs per document type and find existing records.
019d760fst.list domains
Verifies which crawler limits are mapped inside specific index scopes across your platform.
019d760fst.list engines
Extracts a list of all active, isolated Elastic indices running within the Headless Swiftype Platform.
019d760fst.post search
Runs raw queries against a specified engine and returns results in a structured JSON format with facets.
019d760fst.post suggest
Provides predictive key suggestions and spelling-tolerant matches without needing to search the main index.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Swiftype, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Listen up. The Swiftype MCP Server lets your AI agent connect straight into your Elastic Search backend. You don't need to open a web dashboard or write complicated API calls; your agent treats this like it’s already integrated, letting you manage the entire data lifecycle right from your prompt environment.
Discovery and Indexing Architecture:
Your agent can start by figuring out what's running under the hood. Use st.list_engines to get a complete list of every isolated Elastic index bound to your account. If you need to know what kind of data is stored, run st.list_doc_types to extract schema blueprints and see exactly what object hierarchies exist in your indexes.
To check the scope limits across your platform, use st.list_domains to verify which crawler limits are mapped inside specific index scopes.
Running Searches and Getting Suggestions:
Need data? You've got two ways to search. For deep dives, fire raw queries into a specified engine using st.post_search. This function returns the results in structured JSON format, including helpful facets that break down the data by category. If you just want quick autocomplete or need spelling-tolerant matches without hitting your main index, use st.post_suggest.
It provides predictive key suggestions instantly.
Managing Document Records (CRUD):
The agent handles all the document bookkeeping. You can check out existing records and get a full dump of stored metadata using st.list_documents, which lets you track IDs per document type and find any record you're looking for. When it’s time to add fresh content, use st.create_documents; this function adds new structured documents in bulk based on a list of attached rules.
If data gets stale or needs cleaning out after validation, run st.delete_documents to permanently remove cached pages and data metrics from the search indices.
Pulling User Behavior Analytics:
You gotta know what people are actually searching for. To see deep arrays listing the most frequent search terms users use, call st.analytics_top_searches. For a clearer picture of conversion, calculate active hit paths by running st.analytics_top_clicks, which identifies the top documents that received the highest number of clicks.
These tools pull real-world data right into your prompt environment without you having to navigate multiple dashboards.
How Swiftype MCP Works
- 1 Append the Swiftype MCP agent to your Vinkius integration variables.
- 2 Set up explicit authorization by binding your
SWIFTYPE_API_KEY(available in the Elastic Admin Profile). - 3 Request a multi-step task, such as: "List my active semantic engines, search the 'help-center' engine for 'billing support', and show me the top 3 analytics clicks this month."
The bottom line is that your AI agent handles all the API calls and data formatting; you just give it a complex instruction.
Who Is Swiftype MCP For?
This server is built for platform engineers, content managers, and data analysts. If you spend time manually clicking through web dashboards to check analytics or run test searches, this tool saves your sanity. It lets your agent do the heavy lifting—the querying and indexing—so you can focus on interpreting the results.
You use it to audit structured map constraints, running bulk CRUD tasks directly in the terminal against JSON REST indices.
You run st.analytics_top_searches and st.analytics_top_clicks to pull real-world conversion data without logging into a web portal.
You test search queries programmatically, using the agent to audit index bugs or validate document structures across multiple domains.
What Changes When You Connect
- You get real-time user behavior metrics. Instead of guessing, run
st.analytics_top_clicksto see exactly which documents drive the most conversions. - Your agent can handle full data lifecycle management. Use
st.create_documentsandst.delete_documentsto programmatically update or clean your index records. - Discovery becomes trivial. Run
st.list_enginesfirst, then usest.list_doc_typesto map out the entire object hierarchy before running any queries. - You bypass web UIs entirely. Instead of clicking through multiple tabs on a search dashboard, your agent executes complex logic using
st.post_searchin one go. - Speed up development by testing suggestions instantly. Use
st.post_suggestto validate autocomplete terms against the engine before building live features.
Real-World Use Cases
Auditing index integrity after a migration
The data reliability analyst needs to ensure all old record types are still visible. They run st.list_doc_types first, then use st.list_documents to dump the metadata for key document types before confirming the structure is intact.
Finding high-value content gaps
A marketing manager wants to know what users are searching for but not finding content on. They run st.analytics_top_searches and cross-reference those results with the current document list from st.list_documents. This identifies immediate content needs.
Building a live, type-safe search feature
A developer builds a new client feature. They use st.post_suggest to handle the front-end autocomplete logic, and then immediately follow up with st.list_engines to know which specific index to run the final query against.
Cleaning up stale or bad data records
After a feature deprecation, an analyst must remove all related content keys. They first identify the relevant documents with st.list_documents and then use st.delete_documents to vaporize them from the index.
The Tradeoffs
Assuming one tool is enough
Trying to find top clicks just by running a general search query via st.post_search.
→
You need two steps: first, run st.list_engines to target the correct index, and then use st.analytics_top_clicks against that specific engine name.
Running CRUD operations without discovery
Attempting to create a document using st.create_documents but forgetting which schema it needs.
→
Always run st.list_doc_types first. This gives you the necessary schema blueprints before injecting new payload structures.
Over-relying on broad search
Using only st.post_search, which can be slow and complex for simple autocomplete.
→
For pure predictive text features, use st.post_suggest. It’s faster and designed specifically to provide bounding keys decoupled from the main index.
When It Fits, When It Doesn't
Use this MCP Server if your task requires deep visibility into structured data or user behavior metrics (e.g., 'Show me what users clicked' or 'List all indexes'). Don't use it if you just need to answer a simple question ('What is the return policy?'); for that, a basic text search tool will suffice.
This server excels when you combine discovery (st.list_engines / st.list_doc_types) with action (CRUD tools) and analysis (st.analytics_*). If your workflow involves multiple steps—like 'Find the engine' -> 'Search it' -> 'Check its analytics'—this is the right tool. Skip it if you only need simple, stateless chat responses.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Swiftype / Elastic App Search. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Checking search performance shouldn't require opening a web dashboard.
Today, checking how well your content performs means logging into the Elastic UI. You click to select the correct index, then navigate to the 'Analytics' tab, manually filter by date range, and finally run multiple queries just to get a list of top-performing documents.
With this MCP server, you tell your agent exactly what you need: "What were the top three searched terms last month?" The agent runs `st.analytics_top_searches` and gives you the raw data structure instantly. No clicks needed.
Swiftype MCP Server: Get structured, actionable data.
Forget having to run multiple API calls manually—one for the list of engines, another for the document types, and a third for the search itself. The agent sequences these steps automatically based on your prompt.
You get full control over complex data flows in one conversation. It's less about running queries and more about commanding a data architecture.
Common Questions About Swiftype MCP
How do I find all the available search engines using st.list_engines? +
Run st.list_engines. This function returns a list of every active, isolated Elastic index that your agent can target for queries.
What is the difference between st.post_search and st.post_suggest? +
st.post_search runs full queries against an engine to retrieve documents (e.g., 'billing support'). st.post_suggest only provides predictive, spelling-tolerant completions before a search happens.
How do I know what kind of data is in my index? +
Use st.list_doc_types. This function extracts the schema blueprints, so you know exactly what object hierarchies are available for querying or creating documents.
Can st.delete_documents delete everything? +
No. You must first use st.list_documents to find the specific document IDs and metadata. The tool requires explicit keys to vaporize data, ensuring you don't lose anything by accident.
What security measures do I need to take when using `st.post_search`? +
You must bind your API key for secure access. The connection requires a validated SWIFTYPE_API_KEY bound within the Vinkius environment variables, ensuring that only authenticated requests can fire raw queries into the specific search engine.
How do I ensure data integrity when calling `st.create_documents`? +
You must provide structured JSON payloads for every document you want to index. The system validates these incoming records against the schema blueprint found using st.list_doc_types, preventing corrupted data from entering your live search index.
Does the platform handle rate limiting when executing multiple queries via `st.post_search`? +
Yes, the underlying API handles throttling errors automatically. If you exceed defined request limits, the agent will return a specific error code (typically 429), allowing your AI client to implement an exponential backoff and retry the query later.
Can I filter the top clicks generated by `st.analytics_top_clicks` by a date range? +
Absolutely, you pass specific start and end dates as parameters when calling st.analytics_top_clicks. This limits the returned data set to only those conversion metrics that fall within your specified operational time window.
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.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
GetBlock (Web3 RPC Provider)
Access 50+ blockchain networks (Ethereum, Solana, Bitcoin) via RPC. Query balances, blocks, and transactions directly from your AI agent.
Netlify Alternative
Modern web development platform — manage sites, deploys, and forms via AI.
Absolute Chronological Timeline Engine
Empower your AI Agent with deterministic chronological precision. Calculate exact ages, compare lifespans, forecast milestones, and track anniversaries — all offline and hallucination-free.
You might also like
Pumble
Bring your AI agent into your Pumble workspace — read channel history, post messages, create channels, and manage team communication seamlessly.
Zenkit
Manage workspaces, lists, and entries via the Zenkit API.
Playground AI
Generate, inpaint, upscale, and transform images using Playground AI's powerful models via natural language.