Running website A/B tests from Claude Code with an MCP server

Описание и применение

AI‑агент может автоматически запускать A/B‑тесты, от подключения сайта до отслеживания конверсий и анализа результатов, используя протокол MCP и сервер Luria.

Задача

Determining if a new landing page converts better than the old one requires a split test on real traffic, which traditionally needs a testing tool, dashboard, and manual monitoring.

Результат

The agent can autonomously run A/B tests, from site setup to conversion tracking and result analysis, with human approval for the first live change and safety guardrails.

Рабочий процесс

• Connect MCP server via claude mcp add --transport http luria https://www.luriart.com/api/mcp
• Add site and install tag with prompt "Add example.com to Luria and install the tag in this repo's layout."
• Create a test on the homepage headline with create_test or auto_test
• Record real conversions server‑side using the /api/v1/conversions endpoint or the agent’s record_conversion tool
• Read test results with get_test and decide on next tests using next_tests/start_next_test

Доказательства

Coding agents can now ship a landing page in minutes. The slow part is finding out whether the new page actually converts better than the old one.

Источник

The Model Context Protocol lets a client (Claude Code, Claude, ChatGPT, Cursor, Codex) call tools on a remote server.

Источник

The agent calls add_site, gets back a one-line script tag and per-framework install notes, edits your layout file, then calls verify_install until the tag reports in.

Источник

Clicks are not wins. If checkout happens on your backend (Stripe, a booking tool), send the conversion server-side: ...

Источник

A change is kept only when it is 95% likely to beat the original after at least 200 visits per version.

Источник