# Amplitude MCP > Amplitude tracks how users move through your product. This MCP connects any AI agent to Amplitude’s full suite of analytics tools so you can query user behavior, monitor feature adoption, and manage behavioral cohorts using natural conversation. You get real-time access to event data, allowing you to analyze complex metrics without leaving your chat. ## Overview - **Category:** marketing-automation - **Price:** Free - **Tags:** product-analytics, user-tracking, event-ingestion, cohort-analysis, data-export ## Description Need to know exactly what users are doing inside the product? This MCP connects your AI agent directly to Amplitude’s powerful analytics engine. Instead of jumping through multiple dashboards and running ad-hoc reports, you just ask your agent a question—like 'How many power users signed up last week?'—and it pulls the answer instantly. You can send high volumes of event data or pull individual user profiles in real time. It even handles complex tasks like identifying behavioral cohorts by uploading simple lists of IDs. This means you don't have to wait for a dedicated data analyst to run reports; your agent does the heavy lifting immediately. Connecting through Vinkius gives you access to this entire suite of tools, letting you analyze product performance and user journeys right where you work. ## Tools ### create_annotation Adds important dates or markers to visual charts for better context. ### create_category Establishes a new grouping or classification type (Taxonomy) for your event data. ### create_release Creates a marker to track the specific feature release that caused an observed change in behavior. ### delete_users Submits requests to delete user data, complying with GDPR and CCPA standards. ### export_data Pulls all raw event data in a standardized JSON format for external use. ### get_audit_logs Retrieves detailed administrative activity logs to track who changed what and when. ### get_experiment_variant Checks which version of an experiment a specific user was assigned to. ### get_segmentation Queries the key performance metrics that are typically displayed within Amplitude charts. ### get_user_profile Fetches real-time data and properties associated with a specific user ID. ### group_identify Sets or updates descriptive properties for defined groups of users simultaneously. ### identify_user Updates a user's property details without generating any new event data. ### list_categories Gets a full list of all established event categories (Taxonomy) used in your system. ### list_cohorts Retrieves an exhaustive list of every behavioral cohort currently tracked. ### list_event_properties Lists all available event properties that capture specific details about user actions. ### list_event_types Retrieves a list of every distinct type of action (event) recorded by users. ### list_user_properties Lists all available properties that describe the user themselves, regardless of their actions. ### map_user Links one unique user ID to another global identifier across different systems. ### send_attribution Sends data detailing where and how a user was acquired from external partners. ### send_batch_events Transmits large volumes of collected event data to Amplitude efficiently in one go. ### send_events Sends a structured array of events directly to the platform using standard API format. ### upload_cohort Creates a new behavioral cohort by accepting and processing a static list of user IDs. ## Prompt Examples **Prompt:** ``` Export raw event data from 2023100100 to 2023100123. ``` **Response:** ``` I've initiated the data export for that range. The raw event data is being retrieved from Amplitude's servers now. ``` **Prompt:** ``` Get the profile and cohort memberships for user ID 'user_123'. ``` **Response:** ``` Fetching profile for 'user_123'... This user is part of 3 cohorts: 'Power Users', 'Early Adopters', and 'Beta Testers'. Their last active property was 'premium_plan: true'. ``` **Prompt:** ``` List all behavioral cohorts in this project. ``` **Response:** ``` I found 12 behavioral cohorts. Some notable ones include 'Dormant Users' (ID: 5a2b), 'Weekly Actives' (ID: 9f1c), and 'New Signups' (ID: 3e4d). ``` ## Capabilities ### Track User Movements Send single or massive batches of event data directly into Amplitude from your agent. ### Fetch User Details Retrieve current user properties, activity status, and recommendation details for any specific ID. ### Analyze Metrics Query complex segmentation metrics or export raw event data to run deep custom analyses outside the platform. ### Manage User Groups (Cohorts) List existing behavioral groups or create brand-new cohorts by providing a list of user IDs. ### Maintain Data Structure Get lists of all available event categories, properties, and types to ensure your data taxonomy stays clean. ## Use Cases ### Debugging a Feature Drop A Growth Engineer notices conversion rates dropped after the last deployment. They ask their agent to 'Check user properties for users who used Feature X but didn't proceed.' The agent uses `get_user_profile` and `get_segmentation` to pinpoint exactly which group is failing, saving hours of manual dashboard investigation. ### Analyzing a Campaign Winner A Product Manager wants to know the core behaviors of users who signed up via a specific paid channel. They ask their agent to 'Find and create a cohort for all users acquired through Google Ads.' The agent uses `upload_cohort` immediately, allowing the team to start studying that group's habits. ### Preparing Data for Offline Analysis A Data Analyst needs historical data for a quarterly deep dive but doesn't want to run it in Amplitude. They ask their agent to 'Export all raw event data for the last 30 days.' The agent uses `export_data` and delivers a clean, ready-to-use JSON file. ### Maintaining Data Integrity A team is rolling out a new product line. Instead of manually updating documentation, the Product Manager asks their agent to 'List all event properties related to payment.' The agent uses `list_event_properties`, ensuring nothing critical was missed before launch. ## Benefits - Stop leaving the chat window. Instead of opening Amplitude, running a query, copying data, and pasting it somewhere else, you can ask your agent to perform complex tasks like fetching user profiles or listing all behavioral cohorts directly in your conversational workspace. - Manage critical data hygiene without manual effort. You can use tools like `create_category` or `list_event_types` to verify your event taxonomy on the fly and ensure new features are tracked correctly from day one. - Scale your analytics workflow with batch processing. Use `send_batch_events` to offload massive amounts of data, ensuring that even high-volume usage doesn't slow down your development cycle. - Improve compliance and governance by managing user data requests through the agent. You can use `delete_users` when a customer invokes their right to erasure, automating tedious legal compliance tasks. - Deepen insights with structured querying. By calling `get_segmentation`, you bypass surface-level metrics and query complex performance indicators needed for true product optimization. ## How It Works The bottom line is that instead of navigating an analytics platform, you just talk to your AI client about the data you need. 1. First, subscribe to this MCP and provide your Amplitude API Key and Secret Key credentials. 2. Next, tell your AI client the task you need done—for instance, 'Show me the profile for user X' or 'Export all events from last week'. 3. Finally, your agent uses the tools available in this MCP to fetch the data or perform the action, and it presents the results back to you right in the chat. ## Frequently Asked Questions **How do I export raw data using Amplitude MCP?** You use the `export_data` tool. Just tell your agent the date range or event criteria, and it will initiate the retrieval of all raw event data in a structured JSON format for you to download. **What is the difference between identify_user and get_user_profile?** `get_user_profile` pulls existing information about a user. `identify_user` allows your agent to update properties—like changing a user's status from 'Trial' to 'Paid'—without actually sending any new event data. **Can I use Amplitude MCP to manage compliance?** Yes. You can trigger the `delete_users` tool through your agent, which handles submitting the necessary requests to delete user data according to GDPR or CCPA rules. **How many different types of events can I track with Amplitude MCP?** You can check all available event types by using `list_event_types`. This tool gives you a comprehensive list of every action your users are currently recording, helping keep your taxonomy clean. **Does this MCP help me test new features?** Yes. You can use the agent to check experiment variant assignments with `get_experiment_variant`, allowing you to see exactly which version of a feature different user groups were exposed to.