Integrations

Mirrored from the OSS repo

This page is auto-synced from docs/integrations.md in github.com/logly/mureo. Edits belong in the OSS repo — this site is a presentation layer.

External Tool Integration Guide

mureo’s core value is orchestration — knowing what to do, when, and why across multiple marketing platforms. mureo includes built-in integrations for Google Ads, Meta Ads, and Google Search Console. For analytics and CRM data that mureo does not cover directly, you can connect third-party MCP servers alongside mureo in the same client. This guide explains how.

How It Works

MCP clients (Claude Code, Cursor, Codex, Gemini, Claude Desktop) can connect to multiple MCP servers simultaneously. Each server exposes its own set of tools. When mureo’s workflow commands run, the AI agent discovers all configured platforms at runtime and adapts its behavior accordingly — calling tools from any connected MCP server in the same session.

Platform Discovery

Every mureo workflow command follows this pattern:

1. Read STATE.json platforms dict to find configured ad platforms
2. Check for built-in data sources (Search Console credentials)
3. Probe for external MCP availability (GA4, CRM) by checking tool namespaces
4. Execute operations across all discovered platforms — no hardcoded assumptions

Key principle: No mureo command requires any specific platform. Commands adapt to whatever is configured. At minimum, one ad platform is needed. All additional data sources (Search Console, GA4, CRM) are additive, never blocking.

GA4 (Google Analytics 4)

Status: Official MCP — first-class mureo support

Google publishes an official Analytics MCP (analytics-mcp on PyPI, repo github.com/googleanalytics/google-analytics-mcp). mureo ships it as an official provider (ga4-official), so you don’t wire it by hand — add it from mureo configure (both the dashboard and the setup wizard list GA4) or with the CLI. It provides read-only access to the GA4 Reporting and Admin APIs.

Configuration

Add the provider with the CLI:

mureo providers add ga4-official

mureo installs the official server via pipx (pipx install analytics-mcp) and registers it in your MCP client config alongside mureo — no .mcp.json hand-editing required. mureo configure walks you through the same step from the dashboard or the setup wizard.

What GA4 Data Adds to mureo Workflows

mureo Command	GA4 Data Value
/daily-check	LP conversion rates, bounce rates, session quality — correlate with ad performance
/search-term-cleanup	LP bounce rates for keyword quality signals
/creative-refresh	LP engagement metrics (time on page, scroll depth) to inform creative direction
/goal-review	Website conversion data for holistic goal tracking
/budget-rebalance	Conversion quality by traffic source to inform allocation decisions
/competitive-scan	Brand/direct traffic trends — competitor mind share signals
/rescue	Site-side vs platform-side diagnosis before making ad changes
/tracking-health	Ads-reported conversions vs GA4 conversions cross-check (>20% divergence flagged)
/weekly-report	Website-level metrics for holistic cross-platform reporting
/onboard	Site conversion metric baseline
/sync-state	Connectivity verification

Authentication

The official GA4 MCP authenticates with a Google Cloud service-account JSON (read-only). mureo providers add ga4-official prompts for the two required environment variables:

• GOOGLE_APPLICATION_CREDENTIALS — path to a service-account JSON key. The service account must be granted access to the GA4 property and the key’s scope must include https://www.googleapis.com/auth/analytics.readonly.
• GOOGLE_PROJECT_ID — the Google Cloud project id.

The GA4 property id is passed per request, not via an environment variable. These credentials are separate from mureo’s Google Ads credentials.

Google Search Console

Status: Built into mureo

Search Console is built into mureo as a first-party integration. It reuses the same Google OAuth2 credentials as Google Ads — no additional authentication or configuration is required.

What It Provides

• Organic vs paid keyword overlap — identify keywords where you rank organically and can reduce paid spend
• SEO/SEM coordination — adjust bidding strategy based on organic ranking changes
• Search appearance data — understand how your pages appear in organic results alongside paid ads
• Indexing status — inspect URLs for indexing issues via the URL Inspection API
• Sitemap management — list and submit sitemaps

Tools

Tool	Description
search_console_sites_list	List verified sites
search_console_sites_get	Get site details
search_console_analytics_query	Query search analytics data
search_console_analytics_top_queries	Get top search queries
search_console_analytics_top_pages	Get top pages by clicks/impressions
search_console_analytics_device_breakdown	Get performance breakdown by device
search_console_analytics_compare_periods	Compare search performance across time periods
search_console_sitemaps_list	List sitemaps for a site
search_console_sitemaps_submit	Submit a sitemap
search_console_url_inspection_inspect	Inspect a URL for indexing status

Workflow Integration

mureo ships 20 workflow commands; the ones that draw on Search Console data are summarized below. Key use cases:

Command	Search Console Value
/daily-check	Organic ranking drops needing paid coverage
/search-term-cleanup	Paid/organic keyword overlap matrix — reduce paid spend on strong organic terms
/competitive-scan	Organic competitive landscape alongside paid auction insights
/creative-refresh	Top organic queries as ad copy inspiration
/budget-rebalance	Organic coverage data to inform paid budget allocation
/rescue	Identify terms better served by organic instead of paid
/goal-review	Organic metrics for SEO-related goals
/weekly-report	Organic trend summary (WoW changes)
/monthly-report	Organic search trend (clicks/impressions/CTR MoM) and paid/organic overlap shifts
/onboard	Organic baseline establishment
/sync-state	Site access verification

CRM / Marketing Automation (HubSpot, Salesforce)

Status: Community MCPs available

Community-maintained MCP servers exist for HubSpot and Salesforce. Quality and completeness vary.

What They Would Add

• Lead quality tracking — connect ad campaigns to downstream lead quality (MQL/SQL rates)
• LTV data — inform /goal-review and /budget-rebalance with actual customer lifetime value
• Pipeline attribution — map ad spend to revenue pipeline for B2B accounts
• Audience sync — verify that ad targeting audiences match CRM segments

Configuration Example

{
  "mcpServers": {
    "mureo": {
      "command": "python",
      "args": ["-m", "mureo.mcp"]
    },
    "hubspot": {
      "command": "npx",
      "args": ["@hubspot/mcp-server"],
      "env": {
        "HUBSPOT_ACCESS_TOKEN": "your-token-here"
      }
    }
  }
}

Note: The HubSpot package name above is a placeholder. Check the official HubSpot or community MCP documentation for the correct package name and configuration.

TikTok Ads

Status: Official MCP — first-class mureo support

TikTok publishes an official TikTok for Business MCP Server (a hosted HTTP MCP). mureo ships it as an official provider (tiktok-ads-official), so you don’t wire it by hand — add it from mureo configure (both the dashboard and the setup wizard list TikTok) or with the CLI. It offers read + write across campaign management, reporting, catalogs, creatives, audiences, and Business Center.

mureo points at TikTok’s Progressive Disclosure endpoint https://business-api.tiktok.com/open_mcp/tt-ads-mcp-layer — ~40 core tools load up front and the rest are discovered on demand — rather than the ~400-tool full-surface (tt-ads-mcp-flat) endpoint, to keep the exposed tool count manageable.

Configuration

TikTok’s MCP supports OAuth Dynamic Client Registration, so it can be added directly as a remote HTTP MCP server and authenticated in the browser on first connect — no developer token, API key, or .mcp.json hand-editing required:

claude mcp add --transport http tiktok-ads \
  https://business-api.tiktok.com/open_mcp/tt-ads-mcp-layer

Then run /mcp in Claude Code, select tiktok-ads, choose Authenticate, and sign in with your TikTok for Business account. mureo configure walks you through the same steps and reflects the connected state on the dashboard.

Note: The official TikTok MCP is in beta; its tool surface may change. On Claude Desktop, add it via Settings → Connectors → Add custom connector with the URL above; on Codex it is not yet wired.

Authentication

Interactive browser OAuth via your TikTok for Business account on first connect. mureo never stores or handles a TikTok token — Claude Code (or a Claude.ai custom connector) brokers the sign-in. These credentials are separate from your other platform credentials.

Future Platforms

The following platforms are planned for integration as their official or community MCP servers mature:

Platform	Status	Expected Value
LinkedIn Ads	Planned	B2B audience targeting coordination, ABM campaign alignment
Amazon Ads	Planned	E-commerce ad spend coordination, product-level ROAS
Microsoft Ads	Planned	Search campaign coordination alongside Google Ads

Pattern: As official MCPs become available from these platforms, add them to your .mcp.json alongside mureo. mureo’s workflow commands will incorporate the additional data opportunistically — no code changes required.

Best Practices

Start with mureo alone

You do not need any external MCP to use mureo effectively. The workflow commands work fully with Google Ads and Meta Ads data alone. Add external MCPs only when you need the additional data.

One server per platform

Each MCP server should cover one platform. Do not try to combine multiple platforms into a single custom MCP server. The MCP protocol is designed for multiple specialized servers working together.

Credential isolation

Each MCP server manages its own credentials independently. mureo credentials (~/.mureo/credentials.json) are never shared with other MCP servers.

Validate before trusting

When an external MCP returns data, the AI agent should cross-reference it with mureo’s own data where possible. For example, compare GA4 conversion counts with Google Ads conversion counts to identify tracking discrepancies.