MCP documentationBahasa Indonesia

Connect Klik Flow to your AI app.

Klik Flow MCP guides: how to connect Claude, ChatGPT, Codex, Cursor and other AI agents to your business data, the permissions they use, the limits, and how it stays secure.

Getting started

What MCP is in Klik Flow, the package and permission requirements, and the first three steps to connect an AI app.

Open guide →Claude Code

One terminal command, then sign in to Klik Flow from the /mcp menu.

Open guide →Claude.ai, Desktop, mobile

Add it once as a custom connector in your Claude account; Desktop and mobile use it too.

Open guide →ChatGPT

Add it as a custom MCP server with OAuth sign-in.

Open guide →Codex

Add the server with codex mcp add, then sign in with codex mcp login.

Open guide →Cursor

Add it in ~/.cursor/mcp.json, then sign in from Cursor's MCP settings.

Open guide →VS Code

Add it in .vscode/mcp.json or with code --add-mcp.

Open guide →Gemini CLI & Antigravity

One command in Gemini CLI; in Antigravity, add a server with the same address.

Open guide →Windsurf

Add a remote server in mcp_config.json.

Open guide →OpenClaw

Add it with the streamable-http transport; no extra skill needed.

Open guide →Hermes Agent

Add it in ~/.hermes/config.yaml with trust: untrusted, then hermes mcp login.

Open guide →n8n

The MCP Client Tool node with an access key, for automation without a screen.

Open guide →Concepts

How Klik Flow MCP decides what an AI may do, and how the owner stays in control.

Open guide →Security & privacy

Number masking, labelled customer text, Indonesia's personal data protection law, Meta's terms, and reporting vulnerabilities.

Open guide →Limits & help

Call limits per connection, what 401, 403 and 429 mean, and answers to common problems.

Open guide →Tool reference

Every Klik Flow MCP tool, parameter, scope and prompt, generated from the server code.

See the reference →Changelog

What changed in each version of the Klik Flow MCP server.

See the changelog →