AdPlug launches MCP server to connect Google Ads and other platforms to AI

Automating the PPC workflow

AdPlug has launched a hosted MCP server designed to bridge the gap between advertising platforms and AI assistants. MCP stands for Model Context Protocol, which is an open standard that allows AI assistants to interact with external software. By using this standard, the tool lets users connect their live ad accounts to popular AI models like Claude and ChatGPT.

The tool was created by Daniel Popa, a veteran marketer who managed over $100 million in ad spend over 12 years. Popa built the software to solve the repetitive manual tasks he faced while managing large budgets. He noted that while spending high amounts of money isn't the stressful part of the job, the manual work of exporting data every Monday was.

The primary goal of AdPlug is to replace the need for manual exports and complex spreadsheet management. Popa described the tool as a way to remove the 'scaffolding' of marketing work. It allows specialists to ask questions about their accounts and receive answers based on real-time data rather than static files.

Multi-platform data integration

AdPlug provides a centralized way to access data from several different marketing ecosystems simultaneously. The current suite includes connectors for Google Ads, LinkedIn Ads, Microsoft Ads, and Reddit Ads. It also integrates with Google Search Console for organic data and HubSpot for tracking leads, deals, and revenue.

For Google Ads specifically, the tool can perform search term audits, check pacing, and identify wasted spend across multiple accounts. The LinkedIn Ads connector provides data on campaigns, audiences, and creative performance. It also allows for approved edits to be pushed directly through the AI interface.

The Microsoft Ads connector is particularly useful for reporting on Performance Max campaigns and copying existing Google Ads campaigns over to the Microsoft platform. By bringing these different sources into one place, the tool allows marketers to see how paid media interacts with organic search and sales data.

Safety and human oversight

Because the tool handles live budgets, safety was a primary requirement during its development. AdPlug is designed so that the AI cannot make autonomous changes to an account without permission. Users must decide on a platform-by-platform basis what actions the assistant is allowed to perform.

When a user requests a change, the system shows the exact field, the current value, and the proposed new value. The AI then waits for a human to provide final approval before the update is executed. Every single call made by the AI is logged to ensure there is a clear audit trail for the account.

Data privacy is also a core component of the service. AdPlug ensures that ad account data is never used to train any AI models. This is a critical feature for agencies that handle sensitive client information and must comply with strict data protection standards like GDPR.

Availability and future plans

AdPlug is currently available for use within Claude, Claude Code, ChatGPT, and Codex. The company offers a free tier that covers one platform without requiring a credit card. For users who need more extensive access, paid plans begin at $29 per month.

The tool is operated by MEANDATA IT S.R.L. and includes a data processing agreement for enterprise users. It was built by a self-taught marketer who started his career making logos at age 16 and eventually moved into the fast-paced world of PPC advertising.

Looking ahead, the development team plans to expand the platform's capabilities further. A TikTok Ads MCP is currently planned for the next release. This will allow marketers to include TikTok performance data alongside their existing paid media mix in their AI-driven reports.