mobile-growth-mcp
MCP server for mobile growth & user acquisition, connecting LLMs to a curated knowledge base of ad optimization insights and analytical skills for subscription apps.
README
mobile-growth-mcp
MCP server for mobile growth & user acquisition. Connects your LLM to a curated knowledge base of ad optimization insights and a library of analytical skills (campaign audits, fatigue detection, weekly diagnostics). Focused on subscription apps.
For Meta Ads data, use Meta's official Meta Ads MCP / AI connector. The skills here interpret what it returns. This MCP intentionally does not ship Meta API tools — keeping users on the official path.
For Google Ads, this MCP includes direct API integration (your tokens stay on your machine).
Quick Start
Claude Desktop
-
Open the config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
- macOS:
-
Add this to the file (create it if it doesn't exist):
{
"mcpServers": {
"mobile-growth": {
"command": "npx",
"args": ["-y", "mobile-growth-mcp"],
"env": {
"API_KEY": "me_YOUR_KEY"
}
}
}
}
- Replace
me_YOUR_KEYwith the API key from your account admin. - Restart Claude Desktop.
- To analyze Meta data, also install Meta's official Meta Ads MCP — its connector data flows through your LLM into the skills here.
Note: Use the JSON config file, not the "Add custom connector" UI in Claude Desktop. That UI is for remote servers — this MCP runs locally.
Other clients
| Client | Config file |
|---|---|
| Claude Code | .mcp.json in your project root |
| Cursor | .cursor/mcp.json |
| Codex | MCP config in your Codex setup |
Same JSON format as above.
Alternative: CLI args
{
"mcpServers": {
"mobile-growth": {
"command": "npx",
"args": ["-y", "mobile-growth-mcp", "--api-key=me_YOUR_KEY"]
}
}
}
No API key?
The server still starts — you get the Google Ads tools (if configured) and a connection_status tool that explains what's missing and how to fix it.
How It Works
The MCP server runs as a local process on your machine, started by your MCP client (Claude Desktop, Cursor, etc.).
- Knowledge base (remote) — curated expert insights hosted on our server, accessed via your API key
- Google Ads API (local) — calls happen directly from your machine using your tokens
- Meta Ads data — sourced from Meta's official AI connector, not from this MCP
Your machine Our server
+---------------------------+ +------------------+
| MCP client (Claude, etc.) | | |
| | | | Knowledge base |
| v | API | (insights, |
| mobile-growth-mcp --------+--key--->| search, |
| | | | embeddings) |
| | | +------------------+
| v
| Google Ads API (your tokens stay local)
|
| (Meta Ads data: comes via Meta's official AI connector,
| not through this MCP)
+---------------------------+
Your ad platform tokens are used locally and are never sent to our servers.
What You Get
Knowledge base tools — search, browse, and retrieve expert insights on creative strategy, audience targeting, scaling, bid optimization, and more.
Google Ads API tools — pull campaigns, ad groups, assets, performance insights, and network mix directly from your Google Ads account.
Analytical skills (MCP prompts) — KB-grounded methodologies for ad fatigue, weekly performance, creative health, placement audits, audience composition, architecture reviews, campaign comparisons, attribution analysis, and full account audits. The Meta-flavored skills consume data from Meta's official AI connector or pasted CSVs.
Tools
| Tool | Description |
|---|---|
connection_status |
Check KB and Google Ads connection status; shows how to fix issues |
search_insights |
Semantic + keyword hybrid search across curated insights |
list_insights |
Browse all insights with optional filtering |
get_insight |
Full content of a specific insight by slug or ID |
get_google_ads_campaigns |
List Google App Campaigns with status, bid strategy, budgets |
get_google_ad_groups |
List ad groups within Google App Campaigns |
get_google_assets |
List creative assets with slot utilization audit |
get_google_insights |
Performance metrics with network/device breakdowns |
get_google_network_mix |
Detect Search ↔ Display ↔ YouTube traffic shifts |
get_google_asset_fatigue |
Asset-level fatigue detection |
upload_google_image_assets |
Upload image assets to Google Ads |
Report Prompts (Skills)
Each Meta-flavored prompt takes an ad_account_id and walks the LLM through a structured analysis. Data comes from Meta's official AI connector or pasted CSV exports.
| Prompt | Description |
|---|---|
ad-fatigue-report |
Detect creative fatigue with daily granularity |
weekly-performance |
Week-over-week health with Kast diagnostic framework |
creative-performance |
Categorize ads: scaling / promising / fatiguing / dead weight |
audience-composition |
Age × gender heatmap with CPA analysis |
architecture-review |
Campaign structure evaluation |
audit-meta-account |
Comprehensive account audit |
campaign-comparison |
Side-by-side campaign comparison |
placement-audit |
Detailed placement audit with waste quantification |
attribution-analysis |
Conversion quality validation |
google-campaign-health |
Google App Campaign WoW health check |
google-asset-performance |
Google asset categorization by actual CPI |
google-network-audit |
Search/Display/YouTube spend analysis |
google-architecture-review |
Google App Campaign structure audit |
google-bid-strategy |
Google bid strategy + signal volume evaluation |
All reports reference specific knowledge base insight IDs so recommendations are grounded in expert knowledge, not generic advice.
Resources
| URI | Description |
|---|---|
vocabulary://tags |
Topic tags, applies_to tags, and platforms with counts |
instructions://getting-started |
Full guide to tools, reports, and usage |
Development
Setup
git clone https://github.com/kubachour/mobile-growth-mcp.git && cd mobile-growth-mcp
npm install
npm run build
Environment
cp .env.example .env
# Fill in API_KEY (get from admin)
# For Google Ads tools: run `npx mobile-growth-mcp auth google` to set up
# For ingestion/admin: also set SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY
Adding Insights
- Create or edit JSON files in
data/insights/ - Follow the schema in
packages/shared/src/types.ts - Run
npm run ingest - Embeddings are generated automatically server-side
Publishing
cd packages/mcp-server
npm version patch
npm publish --access public
Project Structure
packages/shared/ — Types, Supabase client
packages/mcp-server/ — MCP server (published as mobile-growth-mcp)
src/google/ — Google Ads API client
src/tools/ — All MCP tools (KB + Google Ads)
src/resources/ — MCP resources (vocabulary, instructions)
packages/ingestion/ — CLI to validate & upsert insight JSONs
data/insights/ — Curated insight JSON files
skills/ — Skill markdown files (analytical methodologies)
supabase/migrations/ — SQL migrations
supabase/functions/ — Edge Functions (embed, search)
License
This project uses a dual license:
- Code (everything except
skills/) — MIT License - Skill prompts (
skills/*.md) — CC BY-NC-SA 4.0 — you can read, learn from, and adapt them for non-commercial use, but you can't use them to build a competing commercial product
The curated knowledge base is not included in this repository.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.