Integrations · MCP
Model Context Protocol for AI agents
@senderz/mcp on npm — the only Model Context Protocol server in the messaging market. Any MCP-aware agent can send iMessage and SMS, list conversations, manage templates, and check delivery via typed tool schemas. See /integrations/claude-code for the Claude Code-specific setup guide.
What you get
- • @senderz/mcp on npm — install with npx, no build step.
- • Works with any MCP-aware agent — Claude Code, Cursor, Windsurf, custom agents.
- • 15+ tools exposed — send_message, list_messages, create_template, check_capability, and more.
- • Typed tool schemas — every tool has a JSON schema for parameters and return types. Agents get auto-complete and validation.
Frequently asked questions
MCP vs REST API — when should I use which?
Use MCP when an AI agent (Claude Code, Cursor, or a custom agent) needs to send messages as part of an agentic workflow. Use the REST API when your application code calls senderZ directly. MCP wraps the REST API — the underlying endpoints are identical.
Which AI agents support MCP?
Claude Code, Cursor, Windsurf, and any agent built on the Model Context Protocol specification. The list grows as more tools adopt MCP. @senderz/mcp works with all of them.
Is MCP the same as function calling?
Related but different. Function calling lets an LLM invoke functions you define in a prompt. MCP is a standardized protocol for tools that any compatible agent can discover and use without per-prompt configuration. @senderz/mcp exposes senderZ tools via MCP so agents discover them automatically.
Try the MCP server
One command to install. 14-day free trial for the API key.
claude mcp add senderz -- npx @senderz/mcp --api-key sz_live_xxx