$ loading_
Manage A/B experiments, traffic splits, results, and winning rollouts via MCP.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "Kumiki A/B" yet — see the docs or source repo.
Create an A/B test for the checkout page button copy. Variant A is "Buy Now" and variant B is "Place Order". Start with a 50/50 traffic split and return the experiment ID and configuration summary.
Returns the new experiment ID, variant setup, traffic allocation, and experiment status summary.
Read the latest results for experiment exp_123. If variant B has a significantly higher conversion rate than A, change the traffic split to 20% for A and 80% for B, then summarize the key metric changes.
Outputs a results summary, significance assessment, updated traffic split, and reason for the change.
Check whether experiment exp_123 has met the winner criteria. If it has, apply the winning variant as the default, end the experiment, and generate a short rollout note.
Returns whether the winning variant was rolled out, the final experiment status, and a ready-to-use rollout note.
Run tests programmatically, inspect results, and get test strategy recommendations.
Connects experiment platforms and adds local tools for analysis and skill debugging.
Control Chrome locally or remotely to automate browsing, forms, screenshots, and scripts.
Access ABsmartly to manage experiments, metrics, goals, and team workflows.
Control a live Cloudflare Workers dashboard in real time via MCP tools.
Build, deploy, manage hosted apps, and chat with each app’s persistent AI agent.