An MCP (Model Context Protocol) tool that connects Taboola Advertising API with AI assistants like Claude (in Cursor, Claude Desktop, etc.), allowing you to analyze your campaign data through natural language conversations.
Clone or download this repository and install dependencies
Install dependencies
pip install -r requirements.txt
Set up your environment variables
Create a .env
file in the root directory with your Taboola API credentials:
TABOOLA_CLIENT_ID=your_client_id
TABOOLA_CLIENT_SECRET=your_client_secret
Test the connector
Run the script to verify it connects to Taboola correctly:
python taboola_connector.py
You should see logs indicating successful authentication and account loading.
Add the connector to Cursor’s MCP configuration
Open or create .cursor/mcp.json
in your user directory and add the following entry:
{
"mcpServers": {
"Taboola_Connector": {
"command": "FULL/PATH/TO/YOUR/DIRECTORY/taboola_mcp/.venv/Scripts/python",
"args": [
"FULL/PATH/TO/YOUR/DIRECTORY/taboola_mcp/taboola_connector.py"
]
}
}
}
Restart Cursor
Close and reopen Cursor to load the updated MCP configuration.
Verify the connection
In Cursor’s chat, ask a simple question like “List my Taboola accounts” to verify the connection works.
The connector provides the following tools for interacting with Taboola:
Tool Name | Description | Example Query |
---|---|---|
list_taboola_accounts |
Lists all accessible Taboola accounts | “Show me all my Taboola accounts” |
taboola_account_overview |
Provides a comprehensive account overview | “Give me an overview of account wattfoxgmbh-network” |
get_taboola_account_currency |
Retrieves the account’s currency code | “What currency is my Taboola account using?” |
list_taboola_campaigns |
Lists all campaigns in an account | “List all campaigns in my wattfoxgmbh-network account” |
get_taboola_campaign_summary |
Gets performance summary for campaigns | “Summarize campaign performance for April” |
get_taboola_campaign_details |
Gets detailed info about a specific campaign | “Show me details for campaign 44471496” |
analyze_taboola_performance |
Analyzes campaign performance with detailed metrics | “Analyze my campaign performance from April 1 to April 23” |
get_taboola_creatives |
Lists all creative assets in an account | “Show me all creatives in my Taboola account” |
get_taboola_creative_details |
Gets detailed info about a specific creative | “Show details for creative asset 123456” |
analyze_taboola_creative_performance |
Analyzes creative performance with metrics | “Which creatives performed best last month?” |
get_taboola_ads |
Lists all ads or ads for a specific campaign | “Show me all ads in campaign 44471496” |
download_taboola_creative |
Downloads a creative asset from its URL | “Download the creative at https://example.com/image.jpg” |
You can ask questions in natural language like:
.env
file exists with correct credentials..cursor/mcp.json
is correctly specified.taboola_mcp.log
.TABOOLA_CLIENT_ID
and TABOOLA_CLIENT_SECRET
are correct.The connector communicates with the Taboola API to retrieve campaign data, and exposes this data through MCP tools. This allows AI assistants to query and analyze your advertising data.
The connector follows these steps:
If you’d like to contribute to this project, found a bug or an issue, please open an issue or submit a pull request :)
MIT License
A open-source library enabling AI models to control hardware devices via serial communication using the MCP protocol. Initial support for Raspberry Pi Pico.
A Model Context Protocol (MCP) server that enables AI assistants to perform network scanning operations using NMAP
An MCP server enhances AI responses with real-time search results via Higress ai-search.