Make MCP Server – Run and Manage Automations
Make's official MCP server lets AI clients run active and on-demand scenarios and, when scopes allow, manage Make account resources. Use it to turn Make scenarios into callable AI tools, trigger automations from prompts, and view or modify scenarios, webhooks, data stores, teams, and organizations.
Overview
Make provides an official cloud-hosted MCP server that lets AI systems run scenarios and manage resources in a Make account. It connects MCP clients such as Claude, ChatGPT, and compatible developer tools to Make scenarios, teams, organizations, webhooks, connections, data stores, and other account resources according to the scopes granted during connection.
What the MCP server enables
Make MCP turns active and on-demand Make scenarios into callable tools for AI. When a scenario is configured with inputs and outputs, an AI client can pass structured data into the scenario and receive structured results back. This lets an agent trigger existing automations from natural-language prompts while keeping the execution in Make.
The server also provides management tools, depending on the user’s plan and granted scopes. Documented capabilities include:
- Running active scenarios with on-demand scheduling.
- Retrieving scenario outputs after a timeout using an execution ID when the required read scope is granted.
- Viewing and modifying scenarios and related entities such as connections, webhooks, and data stores.
- Viewing and modifying teams and organizations.
- Using OAuth or MCP token connections, with selected scopes determining which tools are exposed.
Scenario-run tools are available on all plans. Management tools are available on paid plans.
When to use it
Use Make MCP when an AI assistant should trigger or manage Make automation from a chat or coding client. Practical examples include running a lead-enrichment scenario, launching an approval workflow, checking data stores, triggering notification automations, updating scenario configuration, or inviting a member to an organization from an AI-assisted workflow.
It is especially useful when an organization already uses Make scenarios and wants them exposed as reusable AI tools.
Connection and authentication
Make documents a cloud-hosted MCP server with Streamable HTTP and SSE transports. OAuth connections use the global hosted URLs:
https://mcp.make.com
https://mcp.make.com/sse
MCP-token connections use the organization zone in the URL. Stateless Streamable HTTP is Make’s default recommended transport for clients that support it. Token-based connections can place the MCP token in the URL path or in an Authorization: Bearer header, depending on client support.
Key considerations
Scopes determine the available tools. OAuth can restrict access to a specific organization, while MCP token connections require careful control of which scenarios are available as tools. Define scenario inputs, outputs, and detailed scenario descriptions so AI clients know what data to provide and what results to expect. Tool calls have timeout limits; scenario runs may continue in Make after the MCP tool call times out, and outputs can be retrieved later when the necessary read scope is granted. Treat MCP tokens as secrets and avoid placing real tokens in shared files.
Supported Transports
streamable_http
URL: https://mcp.make.com
streamable_http
URL: https://${MAKE_ZONE}/mcp/u/${MAKE_MCP_TOKEN}/stateless
streamable_http
URL: https://${MAKE_ZONE}/mcp/stateless
sse
URL: https://mcp.make.com/sse
sse
URL: https://${MAKE_ZONE}/mcp/u/${MAKE_MCP_TOKEN}/sse
sse
URL: https://${MAKE_ZONE}/mcp/sse
Frequently Asked Questions
- When should an AI agent use the Make MCP server?
- Use it when an agent should run existing Make scenarios, expose active on-demand scenarios as callable AI tools, retrieve scenario outputs, or manage Make resources such as scenarios, webhooks, connections, data stores, teams, and organizations according to granted scopes.
- What does the Make MCP server add to an AI agent's capabilities?
- It gives the agent bidirectional access to Make through MCP, allowing selected scenarios to become AI-callable tools and allowing account-management actions when the connected user or token has the required scopes and plan access.
- What can the AI agent access or manage through Make MCP?
- The agent can run active and on-demand scenarios and, on paid plans with management scopes, view or modify scenarios and related entities such as connections, webhooks, data stores, teams, and organizations. The exact tools exposed depend on the selected OAuth scopes or MCP token permissions.
- How is authentication configured for the Make MCP server?
- Make supports OAuth and MCP token connections. OAuth uses the hosted Make MCP endpoint and consent scopes. MCP-token connections use the organization zone and a generated MCP token, either embedded in the zone URL path or sent as an Authorization: Bearer header when the client supports headers.
- Which transport should be used for the Make MCP server?
- Use Stateless Streamable HTTP when the client supports it, because Make documents it as the default recommended method for connection reliability. Use SSE only for clients that require SSE. For MCP-token connections, use the zone-specific URL and either URL-token or Authorization-header format based on client support.