# Kameleoon MCP > Kameleoon MCP lets your AI agent manage A/B testing and personalization workflows without logging into a dashboard. It gives you direct control over running experiments, checking audience segments, and requesting performance reports for any property or site within your account. Use it to automate the monitoring of live tests and analyze complex variations right from your development environment. ## Overview - **Category:** developer-tools - **Price:** Free - **Tags:** ab-testing, personalization, feature-flagging, experimentation, audience-segmentation, conversion-optimization ## Description Your AI client now controls your Kameleoon experimentation platform. This MCP lets agents handle everything from setup to analysis, managing A/B tests, audience segments, and personalization across all your digital properties. You can ask it to list every active experiment or pull the latest performance reports for a specific campaign. If you're used to bouncing between different dashboards just to check status updates, this changes that. Connect via Vinkius and let your agent do the heavy lifting. It reads site metadata, lists variations associated with an ID, and even requests asynchronous results reports so you can analyze experiment performance on the fly. You get full control of your testing lifecycle without ever leaving your chat interface. ## Tools ### create_experiment Creates a new A/B test when given a name and the site ID. ### list_custom_data Retrieves a list of available custom data dimensions for use in testing. ### list_experiments Lists all experiments within Kameleoon, allowing you to monitor various campaign statuses. ### get_experiment Fetches detailed information about a single, specific experiment. ### get_site Retrieves the full details for one particular site property. ### get_experiment_results Requests an asynchronous report containing the latest performance metrics for a finished experiment. ### list_targeting_rules Lists all predefined rules used to segment and target specific user groups. ### list_segments Provides an overview of all defined audience segments in the account. ### list_sites Gathers a complete list of every site registered across your entire account. ### list_variations Lists all specific A/B variations associated with a given experiment ID. ## Prompt Examples **Prompt:** ``` Show me all active experiments in my Kameleoon account. ``` **Response:** ``` I've fetched your experiments. Currently, you have 4 active experiments, including 'New Checkout Flow (AB)' and 'Home Page Personalization'. Which one would you like to inspect? ``` **Prompt:** ``` What are the variations for experiment ID '12345'? ``` **Response:** ``` For experiment 12345, I found 3 variations: 'Reference' (Original), 'Variant A (Red Button)', and 'Variant B (Blue Button)'. All variants are currently receiving traffic. ``` **Prompt:** ``` List all sites registered in my Kameleoon profile. ``` **Response:** ``` I've retrieved 2 sites from your profile: 'Main E-commerce (Code: SEC-123)' and 'Marketing Blog (Code: BLOG-456)'. ``` ## Capabilities ### Initiate new experiments Creates a brand-new A/B test with a specified name and site ID. ### Check current experiment status Lists all active and past experiments to monitor campaign statuses. ### Inspect variations and segments Retrieves details on specific A/B test variations or defined audience segments used for targeting traffic. ### Identify required sites Queries your entire account to list all registered websites and properties, ensuring accurate environment targeting. ### Analyze performance reports Triggers the retrieval of detailed, asynchronous results reports for any given experiment ID. ## Use Cases ### Debugging a broken personalization flow. The Growth Engineer notices a segment isn't working correctly. They prompt the agent: 'List all segments and check the targeting rules for the checkout page.' The agent uses `list_segments` and `list_targeting_rules`, pinpointing that the rule is too restrictive, saving hours of debugging. ### Getting a performance snapshot pre-meeting. The Product Manager has a meeting about last quarter's testing. They ask: 'What are the results for the homepage test from May?' The agent uses `get_experiment_results` and pulls up the necessary data report, instantly ready to present. ### Verifying site readiness. The team is launching a new feature across multiple properties. A developer asks: 'List all sites in our account.' The agent uses `list_sites`, confirming that both the main e-commerce platform and the microservice landing page are registered and ready for testing. ### Comparing test setups. A PM wants to know what variations exist for a specific experiment ID. They prompt: 'What are the variations for campaign 456?' The agent uses `list_variations`, providing a clean list of all possible test groups instantly. ## Benefits - Monitor live tests instantly. Instead of navigating to the Kameleoon dashboard, you can use `list_experiments` to see all active campaigns and their current status in a single prompt. - Targeting accuracy improves. You can quickly list defined audience segments or check specific targeting rules using `list_segments` and `list_targeting_rules`, ensuring your tests hit the right users every time. - Analyze performance data faster. When an experiment wraps up, you don't waste time setting up reports. Just prompt for results and use `get_experiment_results` to initiate fetching the latest metrics. - Know your environment at a glance. Need to confirm if a site is ready for testing? Use `list_sites` or `get_site` to verify all properties are correctly registered before building an experiment. - Manage complex variations simply. Instead of opening the test details page, you can use `list_variations` to see every single variant tied to an experiment ID, confirming setup instantly. ## How It Works The bottom line is, it lets your AI client talk directly to Kameleoon's backend tools using natural language instructions. 1. First, subscribe to this MCP on Vinkius and enter your unique Kameleoon Client ID and Secret credentials. 2. Next, prompt your AI client with a request, like 'List all active experiments for my checkout flow.' 3. Your agent sends the query through the MCP, receives structured data about experiment status or variations, and reports the actionable findings back to you. ## Frequently Asked Questions **How do I find out what experiments are running using the Kameleoon MCP?** You simply prompt your agent to list all active tests. The system uses `list_experiments` to retrieve a comprehensive overview of every experiment currently set up in your account. **Can I get results for an old test using the Kameleoon MCP?** Yes, you can request performance metrics on past tests. Prompting for reports triggers `get_experiment_results`, which initiates fetching the necessary asynchronous data for analysis. **What if I want to check a site ID before creating an experiment?** Use the agent to fetch details about specific sites. The `get_site` tool retrieves all metadata for one property, confirming its status and readiness before you commit to running a test. **How do I see which user groups are available for testing?** You can list all defined audience segments using the `list_segments` tool. This gives you an immediate overview of every group ready for precise traffic allocation in your A/B tests. **What is the difference between listing variations and getting experiment details with Kameleoon MCP?** Getting experiment details (`get_experiment`) gives general metadata about the whole test. Listing variations (`list_variations`) specifically provides a breakdown of every individual testing group or variant attached to that single ID.