Windsor_ai (Preview)

Windsor.ai provides a unified API to access marketing and analytics data from over 300 platforms including Google Ads, Facebook Ads, LinkedIn Ads, HubSpot, Shopify, TikTok, and more. This connector allows you to retrieve, aggregate, and filter cross-channel marketing data.

This connector is available in the following products and regions:

Windsor.ai (Preview)

Windsor.ai provides a unified API to access marketing and analytics data from over 300 platforms including Google Ads, Facebook Ads, LinkedIn Ads, HubSpot, Shopify, TikTok, and more. This connector allows you to retrieve, aggregate, and filter cross-channel marketing data.

Windsor.ai MCP Server

Overview

The Windsor.ai MCP Server enables AI agents and copilots to interact with your marketing and analytics data through the Model Context Protocol (MCP). This connector provides a unified gateway to query data across 60+ connected data sources including ad platforms, analytics tools, CRMs, and eCommerce systems — all through your Windsor.ai account.

Publisher

Windsor.ai

Service / API

Windsor.ai MCP Server

• Host: mcp.windsor.ai
• Base path: /
• Scheme: https

Prerequisites

Before you can use this connector:

1. Create a Windsor.ai account at windsor.ai if you don't have one.
2. Connect at least one data source (e.g., Facebook Ads, Google Ads, Google Analytics) at https://onboard.windsor.ai.

Authentication

This connector uses OAuth 2.0.

Available Tools

The Windsor.ai MCP Server provides the following tools for data discovery and retrieval:

Discovery

• get_connectors: Get available Windsor.ai connectors and connected accounts. This is the first step — call this to see which data sources and account IDs are available.
• get_options: Get available fields, date filters, and options for a specific Windsor.ai connector and accounts. Call this to explore what fields and configuration options are available before querying.
• get_fields: Get detailed information about specific fields of a Windsor.ai connector, including descriptions, types, and tables. Fields with NUMERIC or PERCENT types are metrics; all others are dimensions.

Data Query

• get_data: Retrieve data from a Windsor.ai connector. Provide the connector ID, account IDs, field IDs, date range, and any filters or options.

Getting Started

Step 1: Discover Your Connectors

Call get_connectors to retrieve all available data sources and their connected account IDs:

1. Use get_connectors to list all connected data sources and accounts

Step 2: Explore Available Fields

Use get_options and get_fields to understand what data is available:

2. Use get_options with a connector ID and account IDs to list fields, date filters, and options
3. Use get_fields with a connector ID and specific field IDs to get field metadata (type, description, table)

Step 3: Query Your Data

Call get_data with the connector ID, account IDs, fields, and date range:

4. Use get_data with connector, accounts, fields, and a date range or date_preset to retrieve data

Common Use Cases

• Cross-channel marketing analytics: Query spend, clicks, impressions, and conversions across Facebook Ads, Google Ads, and other platforms
• Campaign performance reporting: Retrieve campaign-level metrics with date filters and breakdowns by dimension
• Field exploration: Discover all available metrics and dimensions for a connector before building queries
• Filtered data retrieval: Use the filters parameter to narrow results by campaign name, spend threshold, and more

Known Limitations

• Connectors mysql, postgresql, redshift, mongodb, snowflake, and big_query require the date_filters parameter when filtering by date.
• Response data is capped at 100,000 characters (~20,000 tokens). Reduce the date range, number of fields, or number of accounts if the limit is exceeded.
• Only read operations are supported; data cannot be written back to connected sources.

Support

• Website: https://windsor.ai
• Documentation: https://windsor.ai/docs
• Support: support@windsor.ai

Throttling Limits

Actions

Throttling Limits

Actions

Invoke MCP Server