Reaudit
Log In →

MCP Server

AI Agent Integration

Connect OpenClaw, Claude Code, Cursor, and other AI agents to Reaudit via MCP — with SKILL.md auto-discovery, MCP Apps, and example workflows.

Reaudit's MCP server lets any AI agent access 180 tools for AI visibility monitoring, SEO audits, content generation, analytics, paid intelligence (including ChatGPT Ads on OpenAI), and publishing. Agents connect via the standard Model Context Protocol (MCP) using SSE or Streamable HTTP transport.
For agents that support SKILL.md auto-discovery (like OpenClaw), Reaudit provides a skill file that teaches the agent all available tools, optimal workflows, and best practices — no manual configuration needed.

Prerequisites

An active Reaudit subscription (Professional, Agency, or Enterprise) and an API key from Dashboard → Tools → MCP Server.

OpenClaw

OpenClaw (openclaw.dev) is an open-source AI agent that supports SKILL.md auto-discovery. Once configured, OpenClaw automatically learns all 180 Reaudit tools and can execute complex multi-step workflows. Add the Reaudit MCP server to your OpenClaw configuration:
json
{
  "mcpServers": {
    "reaudit": {
      "url": "https://mcp.reaudit.io/sse?apiKey=YOUR_API_KEY"
    }
  }
}
OpenClaw reads the SKILL.md file to discover all tools automatically. No further setup is needed — just start prompting:
text
> Check my AI visibility across all platforms and generate a report

OpenClaw automatically runs:
  reaudit: list_projects
  reaudit: get_visibility_score
  reaudit: get_brand_mentions
  reaudit: get_competitor_comparison
  reaudit: create_report
OpenClaw chains tools automatically. Ask complex questions and the agent will figure out which tools to use and in what order.

Claude Code

Claude Code is Anthropic's CLI agent with native MCP support. Add Reaudit as an MCP server with a single terminal command:
bash
claude mcp add reaudit \
  --transport sse \
  "https://mcp.reaudit.io/sse?apiKey=YOUR_API_KEY"
text
> Audit my site, fix SEO issues, and publish an optimized article

Claude Code automatically runs:
  reaudit: list_audits
  reaudit: get_audit_recommendations
  reaudit: generate_content
  reaudit: publish_to_wordpress
Replace YOUR_API_KEY with your actual API key from Dashboard → Tools → MCP Server.

Cursor

Cursor is an AI-first code editor with built-in MCP integration. Go to Cursor Settings → Tools & MCP → New MCP server and paste this into ~/.cursor/mcp.json:
json
{
  "mcpServers": {
    "reaudit": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.reaudit.io/sse?apiKey=YOUR_API_KEY",
        "--transport",
        "sse-only"
      ]
    }
  }
}
Cursor connects to the hosted server through the mcp-remote npx wrapper, so Node.js 18+ must be installed and on your PATH.

MCP Apps

Reaudit also supports interactive MCP Apps that render visual dashboards directly in the Cursor IDE (Cursor 2.6+):
  • Visibility Score — live gauge showing your AI visibility score
  • Citation Tracker — donut chart with citation distribution
  • AI Crawl Monitor — terminal-style log of AI bot activity
  • AI Readiness Score — score card for website optimization

Other Agents

Any agent or tool that supports the MCP protocol can connect to Reaudit. Common clients and their setup method:
  • Windsurf — MCP settings (URL)
  • VS Code Copilot — settings.json → mcp.servers
  • GitHub Copilot — MCP config
  • Gemini CLI — MCP config
  • Claude.ai — OAuth (one-click, no API key)
  • Claude Desktop — config file (OAuth)
  • Lovable — MCP integration
  • Replit — MCP integration
  • Bolt — MCP integration
  • v0 — MCP integration
Generic MCP configuration (works with any client):
json
{
  "mcpServers": {
    "reaudit": {
      "url": "https://mcp.reaudit.io/sse?apiKey=YOUR_API_KEY"
    }
  }
}
For detailed per-client instructions, see the Connect Your AI Assistant article.

SKILL.md Reference

SKILL.md is a standard file that AI agents read to discover available tools and workflows. Agents like OpenClaw and Claude Code use this file to understand what Reaudit can do, which tools are available, and how to chain them for common tasks.
yaml
name: reaudit
description: AI visibility monitoring, SEO audits,
  content generation, analytics, and publishing
  via 180 MCP tools

# Tool Categories (180 total)
# - AI Visibility (3): brand mentions, scores, competitor comparison
# - SEO Audits (5): site audits, recommendations, alerts, agent-readiness scan
# - Content Generation (9): generate, edit, translate, optimize
# - GTM Strategy (9): strategies, steps, content items
# - Analytics Hub (9): GA4, Search Console, Bing, Clarity (incl. paid metric alerts)
# - Paid Intelligence (18): ad creatives, campaigns, ROI, ChatGPT Ads (OpenAI) CRUD + insights
# - Publishing (5): WordPress, React/webhook, calendar
# - Social Media (5): generate, schedule, publish posts
# - Prompt Tracking (7): topics, prompts, AI responses
# - Site Tracking (5): Cloudflare, AI referrals, Webflow, Wix
# - Competitors (3): track, add, analyze
# - Content Radar (4): detect, inspect, analyze, scan competitor content
# - Skills (1): list reusable agent instruction sets
# - Reddit Monitoring (3): monitors, leads, discussions
# - Indexing (3): connections, URL submission, sync
# - Sources & Outreach (4): citations, authors, opportunities
# - Reports & Grids (11): reports, action grids, agent analytics
# - Account & Budget (9): projects, settings, usage, budget
# - Analytics Query (3): flexible data queries
# - Revenue Attribution (1): unified multi-source attribution

# Recommended Workflows
# 1. Visibility check: get_visibility_score → get_brand_mentions → get_competitor_comparison
# 2. Content pipeline: generate_content → translate_content → publish_to_wordpress
# 3. Full audit: list_audits → get_audit_recommendations → create_strategy
# 4. Analytics: get_analytics_hub → get_ga4_analytics → create_report
# 5. Revenue attribution: get_revenue_attribution (merges GA4 + Cloudflare + Stripe)

How Agents Use SKILL.md

  • Auto-discovery — the agent reads the file once and learns all capabilities
  • Tool selection — when you ask a question, the agent matches it to the right tools
  • Workflow chaining — the agent follows recommended workflows for multi-step tasks
  • Best practices — the agent uses optimal parameters and configurations
You do not need to create or modify the SKILL.md file. Reaudit provides it automatically and keeps it updated as new tools are added.

Example Workflows

These are common multi-step workflows that agents can execute from a single prompt. Copy any prompt into your connected agent.

Full Visibility Report

Prompt: "Generate a complete AI visibility report comparing my brand against competitors across all platforms"
Tools: get_visibility_score → get_brand_mentions → get_competitor_comparison → create_report

Content Pipeline

Prompt: "Write an SEO-optimized article about AI search trends, translate to Greek, and publish to WordPress and social media"
Tools: generate_content → translate_content → publish_to_wordpress → generate_social_posts → schedule_social_post

Audit & Strategy

Prompt: "Run a full site audit, prioritize the critical issues, and generate a strategy to address them"
Tools: list_audits → get_audit_recommendations → create_strategy → generate_strategy_step

Analytics Dashboard

Prompt: "Pull all my analytics from GA4, Search Console, and Bing into a unified summary with alerts"
Tools: get_analytics_hub → get_ga4_analytics → get_bing_search_performance → list_analytics_alerts

Revenue Attribution

Prompt: "Show me which AI platforms and referrers are driving revenue, with a breakdown of period sales and attributed revenue"
Tools: get_revenue_attribution

FAQ

What is the difference between MCP and SKILL.md?

MCP (Model Context Protocol) is the transport protocol that connects your agent to Reaudit's server. SKILL.md is a documentation file that teaches agents about available tools and workflows. MCP is required for all agents; SKILL.md provides an extra layer of intelligence for agents that support it.

Do all agents support SKILL.md?

Not all agents support SKILL.md yet. OpenClaw has built-in support. Other agents like Claude Code and Cursor discover tools through the MCP protocol's built-in tool listing. SKILL.md is optional — all 180 tools work without it via standard MCP.

Can I use multiple agents with one API key?

Yes. A single API key works across all agents and platforms simultaneously. Generate API keys from your Dashboard under Tools → MCP Server. For agencies, project-scoped keys let you manage multiple client accounts.

How do agents handle rate limiting?

The MCP server enforces per-project rate limits and token budgets. Agents receive clear error messages when limits are reached. You can monitor usage and set budget limits via the get_budget_status and set_budget_limits tools.

What happens if my agent calls the wrong tool?

MCP tools are designed to be safe — read-only tools like get_visibility_score cannot modify data. Write tools like publish_to_wordpress require explicit parameters. Agents typically confirm destructive actions before executing them.
mcpagentai agentopenclawclaude codecursorskill.mdworkflowsmcp appswindsurfgeminiintegration

Related articles

MCP Server Overview

Connect Your AI Assistant

Capabilities (180 Tools)