PaidSync
Hosted MCP endpoint that lets AI assistants inspect and update live ad accounts across major ad platforms.
What it does
PaidSync is a hosted Model Context Protocol (MCP) server that connects LLMs and AI assistants to advertisers' existing ad, analytics and tracking accounts so agents can discover tools and execute supported account actions. It is designed for PPC specialists, media buyers, agencies and teams that want programmatic agent access to multiple ad platforms without building custom integrations. MCP endpoint and tool discovery PaidSync exposes an MCP discovery surface so an assistant can list available tools, endpoints and account routing options. The MCP interface documents the actions an agent can call, enabling assistants like ChatGPT, Claude or Gemini to discover capabilities for a connected account before taking action. Multi-platform ad integrations The service provides direct connections to major ad platforms including Google Ads, Meta (Facebook & Instagram), LinkedIn Ads, TikTok Ads and additional channels (X, Snapchat, Reddit, Pinterest and Microsoft/Bing are listed in documentation). Integrations use each platform's OAuth flows or equivalent agency tokens. Read/write campaign actions PaidSync supports programmatic read and write operations shown in its docs and example repos: creating and editing campaigns, adjusting budgets and bids, uploading assets, and routing manager-account (MCC/MCM) operations. The product surface is intended to let agents execute changes rather than only generate recommendations. Authentication, logging and control model Connections use the underlying ad platforms' authentication (OAuth, system user tokens, API keys) and PaidSync documents how credentials and logs are handled. By design the platform grants execution privileges to connected agents, so teams must manage authentication and monitoring according to their operational policies. Workflows and common uses Typical workflows described by the vendor and example code include connecting platform accounts via OAuth, pointing an LLM at the MCP endpoint, and asking the assistant to run audits, reallocate budgets, create campaigns, test creatives or fix tracking through Tag Manager and GA4 integrations. GitHub examples demonstrate Google Ads MCP implementations and usage patterns.