Make MCP. Audit workflows and debug automation from chat.

Listen up. This MCP Server connects your AI client directly to your Make account, letting you treat your whole automation setup—the triggers, the data flows, the connections—like one big database you can talk to. You don't gotta open a browser and click through menus just to check if something broke or what data got lost.

It’s built for auditors and debuggers. You can audit complex workflows and inspect infrastructure details using plain language chat. It turns difficult-to-access automation architecture into natural conversation.

Mapping Your Entire Scope

You've gotta know your playing field first. If you need to see what parts of your business are connected, you use list_connections() to get a list of every single active API link tied to the account—whether it’s Google Sheets or Slack. For security audits, this is key because you can verify who's connected and how deep those permissions go.

If your company manages multiple departments, you first call list_organizations() to pull all distinct workspaces associated with your credentials. From there, you can pinpoint the exact organizational scope you need by calling list_teams(), which returns a list of teams belonging to a specific organization ID.

Tracking and Inspecting Scenarios

Want to know what workflows exist? You call list_scenarios() to get a manifest of every single managed scenario currently running in that scope. If you need the full deep dive on one specific workflow, you pass it an ID to get_scenario(id). This function pulls the complete structural details of that scenario, showing module mappings and trigger settings—it’s like seeing the blueprint for how the whole thing is supposed to run.

When a flow fails, you don't wanna guess. You use list_scenario_logs(id) to pull detailed execution logs for any specific Make scenario run. This lets you pinpoint exactly when an automation broke, what data it was trying to process right before the failure, and the raw error payload. It’s your primary tool for figuring out why something went wrong.

Data Visibility and Infrastructure Checks

The workflows rely on stored information, too. You can use list_data_stores() to see a list of internal Make key-value data tables available in the workspace. These are the persistent data stores that multiple workflows pull from or write to, so you can inspect what historical operational info is sitting there.

This server lets your AI client handle all the heavy lifting: listing every scenario, checking team memberships by organization ID, getting structural details for one workflow, finding out which connections exist, and pulling the specific logs needed to fix it. You don't gotta jump through multiple dashboards; you just ask your agent what you need, and it pulls the data directly.