How the Buron MCP server works
The Buron MCP server gives your AI editor direct access to your team's ad data, knowledge files, dashboards, and launch workflow.
The Buron MCP server lets your AI editor talk to Buron directly. Instead of switching between your editor and the Buron dashboard, the agent in your editor can query your Google Ads data, read and write knowledge files, run dashboards, and file launches, all from the same conversation where you're writing code.
It works with any editor that supports the Model Context Protocol: Claude Code, Cursor, GitHub Copilot, and others. You connect once, sign in through the browser, and the editor handles the rest.
What the MCP server gives your editor
Your editor's agent gets access to 14 tools, grouped into four areas:
Ad data. Run raw GAQL queries against your connected Google Ads account, execute saved queries, or run a named dashboard for a date range. The same data you see in the Buron dashboard is available as a tool call.
Knowledge files. Read, write, list, search, and pattern-match files in your team's knowledge layer. Everything your team has captured, from launch sources to strategy docs, is accessible from the editor.
Dashboards. List every dashboard available to your team (default and custom) and run any of them for a specific date range. The agent gets the full panel data back, KPIs, charts, and detail tables, without you writing a single query.
Editor skills. List and install Buron skills like /launch and /setup-google-ads-tracking directly into your project. The MCP server can serve the latest skill templates so your editor always has the current version.
How it relates to the CLI
The CLI and the MCP server are two ways to do the same things. The CLI works from your terminal or a CI job. The MCP server works from inside an AI conversation. Both talk to the same Buron API and the same team data.
If you've already set up the CLI with buron setup, the MCP server gives your editor's agent a richer, more direct connection. It doesn't replace the CLI. Some workflows, like CI automation and buron setup-ci, still run through the CLI.
Stateless by design
Every request to the MCP server is independent. There's no session to manage, no connection to keep alive, and nothing to clean up. Your editor connects, makes a tool call, gets a response, and moves on. This means you can close and reopen your editor without losing anything.
Was this page helpful?