Keen MCP. Run complex analytics and stream data via natural language.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Keen MCP Server streams events and runs analytics queries. Connect your AI agent to your Keen.io project to automate data collection and analysis.
Your agent can record custom events, and it can run complex queries—like calculating the sum, average, or total count of properties—all using natural language.
It gives you visibility into your event streams and data metrics without writing code.
What your AI agents can do
Average property
Calculates the average value for a specified property in your event data.
Count events
Counts the total number of recorded events within a specific collection.
Count unique
Counts how many unique values exist for a specified property.
Run calculations like sums, averages, and total counts on specific properties within your event collections.
Send custom, structured event data to your project's collections instantly, logging every action.
List all existing event collections, saved query definitions, and cached datasets to map your data footprint.
Find all unique values for a specific property or count how many unique entries exist across a collection.
Fetch configuration details and metadata for the entire Keen project.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Keen MCP Server: 10 Tools for Data Aggregation
Use these tools to run queries, count, sum, and average event data directly through your AI agent's conversational interface.
019d75c0average property
Calculates the average value for a specified property in your event data.
019d75c0count events
Counts the total number of recorded events within a specific collection.
019d75c0count unique
Counts how many unique values exist for a specified property.
019d75c0get project details
Retrieves the configuration and metadata for your entire Keen project.
019d75c0list collections
Lists every event collection available within your project.
019d75c0list datasets
Shows a list of cached datasets available for querying.
019d75c0list saved queries
Retrieves a list of all pre-defined, saved analytical queries.
019d75c0record event
Sends a single, custom event record to a specified collection in your project.
019d75c0select unique
Lists all unique values found for a specified property.
019d75c0sum property
Calculates the total sum of numeric values for a specified property.
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 Keen, 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
Yo, hook up your AI agent to your Keen.io project. This MCP server lets your agent record custom events and run complex analytics queries right off natural language. It takes raw event data and turns it into usable metrics for product analysis and data validation.
Stream and Log Events
Send custom, structured event data to a specified collection using record_event. Your agent logs every action this way.
Calculate Aggregate Metrics
Run calculations like sums, averages, and total counts on specific properties in your event data. You can get the total sum of values with sum_property, find the average value for a property using average_property, and count the total number of recorded events in a collection with count_events.
Discover Data Structure
List every event collection in your project using list_collections. You can also show a list of cached datasets with list_datasets or pull up all your pre-defined, saved analytical queries using list_saved_queries.
Identify Unique Values
Find all unique values for a specific property with select_unique, or count how many unique entries exist across a collection using count_unique.
Retrieve Project Metadata
Fetch configuration details and metadata for your entire Keen project by calling get_project_details.
Your agent doesn't need to write code. It just tells the server what it needs, and the server handles the data crunching.
How Keen MCP Works
- 1 Subscribe to the Keen MCP Server and provide your Keen Project ID and Master Key.
- 2 Tell your AI agent what data you need (e.g., 'What was the total count of page views?').
- 3 The agent invokes the correct tool, which runs the query against Keen.io and returns the final metric to your agent.
The bottom line is, your agent runs analytics queries and streams data through natural language commands, without you touching a database interface.
Who Is Keen MCP For?
This is for Data Analysts and Product Managers who need to test hypotheses or track business metrics immediately. It's for the Product Analyst who's tired of asking a data engineer to write a simple query, or the Data Engineer who needs to quickly validate a new event schema before deploying code.
Runs ad-hoc metrics and aggregation queries on event data without writing code or building a dashboard.
Quickly tests event recording and verifies collection schemas by asking the agent, rather than running boilerplate setup scripts.
Integrates event streaming checks and project status monitoring into a development workflow to ensure data integrity.
What Changes When You Connect
- Run ad-hoc metrics instantly. Need to know the average cart value? Use
average_propertyto calculate it right away, without writing a single SQL line. - See every event log. The
record_eventtool lets you instantly send custom event data to any collection, letting you verify your event schema on the fly. - Avoid deep dashboard diving. Use
list_collectionsto see every event type you've captured, andlist_saved_queriesto see what's already calculated. - Understand data distribution. When you use
select_uniqueorcount_unique, you don't just get a number—you see the actual list of unique values, which is crucial for debugging. - Get full project status.
get_project_detailspulls the project's metadata and configuration, giving your agent a complete view of the environment.
Real-World Use Cases
Validating a new checkout flow
A Product Analyst needs to check if the new checkout page is logging the 'payment_method' field correctly. They ask their agent, and the agent uses record_event to send test data. Then, they use count_events to confirm the data appeared in the 'checkout' collection, validating the entire flow.
Calculating weekly revenue trends
A BI Developer wants to track total revenue for the last quarter. They ask the agent to sum the 'price' property. The agent uses sum_property and list_saved_queries to pull the result, giving them the aggregate metric without manual data exports.
Auditing data completeness
A Data Engineer needs to confirm that a critical user property, like 'user_country', is actually being logged. They use list_collections to find the correct data set, and then select_unique to see if any null values or missing data points exist.
Comparing product popularity
A Product Analyst wants to know which products were viewed the most. They ask the agent to count the total events for 'product_view' and then use average_property to see the average view time across all records.
The Tradeoffs
Assuming a single tool does everything
Trying to calculate the sum of revenue and then the average user count in one command. The agent might fail or only run the first command.
→
Break it into steps. First, use sum_property for the revenue total. Second, use count_events on the relevant collection. Then, chain the results together in your prompt to get the full analysis.
Manually querying data outside the agent
Exporting raw data from the web UI, then trying to upload it to a spreadsheet to calculate the average. This is slow and error-prone.
→
Keep it in the agent. Use average_property directly in your conversation. The server handles the query execution and returns the final, clean metric right back to your workflow.
Forgetting to check data structure first
Asking the agent to calculate the average of a property that might be missing or not numeric. This results in an error and halts the workflow.
→
Always start with list_collections and then get_project_details. This ensures your agent knows the exact schema and which properties are available before attempting calculations.
When It Fits, When It Doesn't
Use this if your core need is simple, ad-hoc data validation or aggregation against an event stream. You need to answer questions like 'How many?' or 'What is the average?' without writing SQL or building a permanent dashboard. The tools are best for flat, semi-structured data sets. Don't use this if you need complex relational algebra, like joining two different collections based on a non-primary key, or if you need to run window functions (e.g., calculating a running total). For those advanced needs, you need a dedicated ETL tool or a custom data warehouse connection. If you just need to check the schema or list available metrics, this server is perfect.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Keen. 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
Sifting through dashboards to find one number is a massive time sink.
Today, finding a single metric—say, the total number of successful signups last month—means jumping between the product analytics tool, the database console, and the BI dashboard. You click through tabs, filter dates, and copy-paste numbers into a summary sheet. It's slow and you're always fighting stale data.
With the Keen MCP Server, you simply ask your agent. You tell it, 'What is the total count of successful signups?' The server executes the `count_events` tool and returns the definitive number instantly. You get the answer, not a link to a dashboard.
Keen MCP Server: Get insights instantly.
Previously, calculating the average revenue per user required writing a complex query that joined user tables with transaction tables. You had to run that query, wait minutes, and then manually interpret the result.
Now, your agent runs `average_property` and gives you the precise number immediately. It works with your existing event streams, making the process fast and conversational. No writing, no waiting.
Common Questions About Keen MCP
How do I use the `record_event` tool with Keen MCP Server? +
You tell your agent to record the event, specifying the collection name and the properties. For example: 'Record a 'purchase' event with price 99.99 and user user_123.' The server sends the data to the specified collection.
Can I calculate the average using `average_property`? +
Yes. You prompt your agent to find the average of a specific property (like 'price') in a collection. The tool executes the calculation and returns the calculated mean.
What is the difference between `count_events` and `count_unique`? +
count_events counts every single log entry in a collection. count_unique counts how many distinct values exist for a property, regardless of how many times that value appeared.
How do I check which properties are available for querying? +
You first use get_project_details to pull the project's metadata. This gives your agent the necessary context about available collections and properties before you attempt any calculation.
How do I use `list_collections` to see what data I can analyze? +
The list_collections tool shows every event stream in your Keen project. This lets you see exactly which data sets are available for counting, summing, or averaging.
What is the difference between `list_saved_queries` and `list_datasets`? +
The list_saved_queries tool shows custom queries you've pre-built. list_datasets shows cached results from those queries, which speeds up repeated analysis.
How do I get project details using `get_project_details`? +
Running get_project_details gives you the core metadata and configuration for your entire Keen project. You use this to verify keys or check overall project health.
Can I use `select_unique` on collections I can't see? +
No. You must first use list_collections to identify the target collection. The select_unique tool requires a valid collection ID to execute.
Can I record a single event using this agent? +
Yes, the record_event tool allows you to send a JSON object representing your event directly to any named collection.
How do I calculate the average value of a property? +
Use the average_property tool by specifying the collection name and the numeric target property you want to average.
Is it possible to see my project's collections? +
Absolutely. The list_collections tool retrieves a list of all event types (collections) currently stored in your Keen project.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Google Search Console
Monitor your website's search performance, fix indexing issues, and manage sitemaps via AI.
Unkey API Management
Manage and verify your user API keys via Unkey — create, revoke, and track usage directly from any AI agent.
SambaNova (AI Inference)
High-speed AI inference for Llama 3, DeepSeek, and MiniMax models via SambaNova's ultra-fast SN40L chips.
You might also like
Dogechain Explorer (Dogechain Block Explorer API)
Automate Dogecoin blockchain analysis via Dogechain — check address balances, inspect transactions, and query block data directly from any AI agent.
Zeplo (Queue & Background Job API)
Schedule, queue, and manage background jobs and webhooks via Zeplo directly from your AI agent.
Heymarket
Text customers from a shared business number with two-way SMS, scheduled messages, and CRM integrations for team communication.