Zwitch MCP

The Zwitch MCP Server lets AI assistants like Claude, Cursor, VS Code, Gemini CLI, and Windsurf work directly with Zwitch's developer surface through the Model Context Protocol (MCP) standard.

It puts Zwitch documentation search and ready-to-paste integration code generation inside your AI assistant — so onboarding to Zwitch happens in your IDE instead of across browser tabs.

Key Capabilities

The server ships 11 tools in two groups:

• Documentation (4 tools) — semantic search, browse, and read across the live Zwitch developer docs.
• Integration Guides (7 tools) — generate production-grade integration code for payment gateway, virtual accounts, beneficiaries, payouts, collections, KYC verification, and webhooks.

Languages supported: nodejs, python, go, php (with optional framework parameter).

See the full list in the Tools Reference.

Why Use It

• Zero setup — connect to the hosted endpoint over HTTP. No Docker, no local processes, no API keys.
• Read-only and safe — every tool either searches docs or generates code. Nothing the assistant can call moves money or mutates state.
• Always-current docs — the server searches the same content as developers.zwitch.io.
• Production-grade snippets — integration guides emit real handlers with env-var loading, error handling, idempotency, and webhook signature verification.

Deployment

Zwitch MCP is a remote, hosted service. Connect your AI assistant to the endpoint — no binary to download, no container to run.

See Remote MCP Server Setup to install in your assistant.

Getting Started

1. Pick your AI assistant (Claude Desktop, Claude Code, Cursor, VS Code, Gemini CLI, or Windsurf).
2. Follow Remote MCP Server Setup.
3. Restart the assistant and try "search the Zwitch docs for ..." or "generate the Node.js payment gateway integration."

Reference

• Use Cases
• MCP Server Setup
• Tools Reference
• FAQs

Support

• Dashboard: app.zwitch.io
• Developer Docs: developers.zwitch.io
• Contact: [email protected]

•