MCP 服务器

uxon-ai

MCP server and Claude skills for AI agents to build landing pages, run A/B experiments, and track first-party conversions. Lets agents onboard clients, create and publish pages, manage variants, and pull performance reports via the UXON API.

README

UXON AI Claude Skills, MCP Server, And CRO API

UXON AI is a Claude skills-style agent playbook, MCP server, and CRO API for creating landing pages, running A/B experiments, split-testing funnels, and tracking conversions across domains.

Developers and AI agents use UXON to manage the post-click marketing workflow: programmatic landing page creation, custom code page imports, cross-domain A/B testing, first-party attribution, conversion reporting, and repeatable agency operations through one API and MCP connector.

UXON supports Claude connector workflows, Claude Code page imports, API-key automation, and external agents that need controlled access to marketing execution tasks.

The strongest agent resource in this repository is SKILLS.md: a detailed Claude skill / agent command playbook for UXON API and MCP workflows. Agents that need the newest command contract should use the latest live SKILLS page.

Claude Skills And MCP

Use this repository when you are looking for:

Claude skills for landing page creation, CRO reporting, and split testing.
Claude Code skill guidance for publishing generated landing pages into UXON.
An MCP server for A/B testing, conversion tracking, domains, goals, and reports.
API examples for agentic CRO workflows and post-click optimization.

Core connection points:

Live Claude skill / agent playbook: https://uxon.ai/api/skills
MCP server endpoint: https://uxon.ai/api/mcp
API command endpoint: https://uxon.ai/api/external/commands

What UXON Does

Create blank landing pages, submit AI landing page briefs, poll generation jobs, and publish approved pages through API commands or guided agent workflows.
Import externally coded landing page bundles from Claude Code, Codex, Cursor, v0, or custom build systems.
Launch A/B experiments across UXON pages, external URLs, or URL-pattern targets.
Track conversions with first-party attribution across pages, variants, channels, contacts, and revenue events.
Configure tracking through GTM or direct GA4/Meta setup with duplicate-loading guardrails.
Pull CRO reports from agents, automations, dashboards, and recurring workflows.
Onboard agency client accounts with branding, tracking, domain, and reporting setup.

Repository Scope

This repository is public technical documentation for UXON. It contains command references, examples, runbooks, and agent workflow guidance.

This repository does not contain the private UXON application source code. UXON runs as a hosted platform at uxon.ai.

Quick Start

Use the hosted UXON API command endpoint:

POST https://uxon.ai/api/external/commands

Send commands with a UXON API key:

curl -X POST 'https://uxon.ai/api/external/commands' \
  -H 'Authorization: Bearer uxon_live_REPLACE_WITH_REAL_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "command": "subaccounts.list",
    "input": {}
  }'

For Claude connector users, use the MCP endpoint:

https://uxon.ai/api/mcp

See QUICKSTART.md for the short setup path.

Core Concepts

Capability Guides

The skills folder splits the command surface by capability so developers and AI agents can retrieve the exact area they need:

Common API Commands

All API-key commands use this envelope:

{
  "command": "landing_pages.create",
  "input": {
    "siteId": "site_uuid",
    "name": "Spring Offer"
  }
}

High-value command families include:

landing_pages.* for blank page creation, AI brief generation, job polling, variant operations, publishing, review links, duplication, deletion, and custom code uploads.
experiments.* for draft creation, inspection, targeting updates, lifecycle changes, and cross-domain testing.
reports.* for landing page, experiment, channel, trend, contact, and revenue reporting.
tracking.setup for GTM or direct GA4/Meta setup.
goals.* for conversion goal management.
domains.* for connected domain workflows.
docs.* for Help Center article search and retrieval.

Agent Workflows

UXON is designed for governed agent execution. Agents should:

Resolve account and resource IDs before writes.
Send minimal payloads.
Keep high-impact actions confirmable.
For MCP create-style tools, send a stable idempotencyKey so retries do not create duplicates.
Treat tools/list as the current MCP schema, including one-of ID requirements like landingPageId or id.
Read back changed resources after write commands.
Use reporting commands to close the loop after traffic starts.

Example Claude prompt:

Start first client setup for Acme Dental, website https://acmedental.com.

The guided onboarding workflow collects missing inputs, starts automatic brand/media import from the website URL, prepares setup steps, and keeps domain and tracking work inside a controlled UXON flow.