make-mcp-reference（make MCP reference 搜索）

Make MCP Server Reference

Technical reference for the Make.com MCP server — enables AI clients to execute scenarios and manage Make accounts.

Connection Methods

Codex Quick Setup

If the Codex plugin installs but the Make MCP server is not present in codex mcp list, add it manually:

codex mcp add make --url https://mcp.make.com
codex mcp login make

After login, reconnect the Codex session or start a new one before testing tools.

OAuth (Default)

Connect via OAuth consent flow. Select organization and scopes during authentication.

Endpoint: https://mcp.make.com

URL variants:

For clients without SSE support, a legacy transport using the Cloudflare mcp-remote proxy wrapper is available: npx -y mcp-remote https://mcp.make.com/sse.

Configuration for Claude Code:

{
  "mcpServers": {
    "make": {
      "type": "http",
      "url": "https://mcp.make.com"
    }
  }
}

Access control: Restrict to specific organizations during OAuth consent. Teams plan or higher enables team-level restrictions.

MCP Token

Generate a token in Make profile → API access tab → Add token.

Endpoint: https://<MAKE_ZONE>/mcp/u/<MCP_TOKEN>/stateless

URL variants:

Configuration for Claude Code:

{
  "mcpServers": {
    "make": {
      "type": "http",
      "url": "https://<MAKE_ZONE>/mcp/u/<MCP_TOKEN>"
    }
  }
}

Replace <MAKE_ZONE> with the organization's hosting zone (e.g., eu1.make.com, eu2.make.com, us1.make.com).

Security: Treat MCP tokens as secrets. Never commit them to version control.

Scopes

Scenario Run Scopes

Allow AI clients to view and run active, on-demand scenarios.

• OAuth scope: "Run your scenarios"
• Token scope: mcp:use
• Available on: All plans

Management Scopes

Allow AI clients to view and modify account contents (scenarios, connections, webhooks, data stores, teams).

• Available on: Paid plans only
• Enable granular control over Make account management

Configuring Scenarios as MCP Tools

For a scenario to appear as an MCP tool:

1. Set scenario to active status
2. Set scheduling to on-demand
3. Select the appropriate scope (mcp:use for tokens, "Run your scenarios" for OAuth)
4. Configure scenario inputs — these become tool parameters
5. Configure scenario outputs — these become tool return values
6. Add a detailed scenario description — strongly recommended to help AI understand the tool's purpose and improve discoverability

Input/output best practices:

• Write clear, descriptive names (AI agents rely on these)
• Add detailed descriptions explaining expected data
• Use specific data types over Any
• Keep execution time under timeout limits

Access Control (Token Auth)

Restrict which scenarios are available via URL query parameters:

Organization level:

?organizationId=<id>

Team level:

?teamId=<id>

Scenario level (single):

?scenarioId=<id>

Multiple scenarios:

?scenarioId[]=<id1>&scenarioId[]=<id2>

Levels are mutually exclusive — cannot combine organization, team, and scenario filters.

Timeouts

When a scenario run exceeds the timeout, the response includes an executionId. The scenario continues running in Make for up to 40 minutes. Use executions_get with that ID to poll for results.

Advanced Configuration

Tool Name Length

Customize maximum tool name length with query parameter:

?maxToolNameLength=<32-160>

Default: 56 characters.

Troubleshooting

Resources

• references/transport-details.md — Detailed transport comparison, URL construction, and zone list
• Make MCP Server docs — Official documentation
• make-scenario-building skill — Scenario construction: routing, filtering, iterations, aggregations, error handling, blueprint construction
• make-module-configuring skill — Module configuration: parameters, connections, mapping, webhooks, data stores