Adonis MCP

Adonis MCP fills a niche for developers building AI‑enabled applications on the AdonisJS framework: it exposes a fully‑featured Model Context Protocol (MCP) server that communicates over Server‑Sent Events (SSE). By doing so, it allows Claude and other MCP‑compliant assistants to discover, invoke, and interact with custom tools, resources, and prompts that live inside an Adonis application. This eliminates the need for separate microservices or complex HTTP APIs, keeping the entire AI workflow in a single, type‑safe codebase.

The server is configured via a simple file where developers can set the SSE endpoint, message routing path, and metadata such as server name and version. Once configured, a single helper method () lets you declaratively register three core MCP entities:

• Resources – lightweight, URL‑based endpoints that return structured data. In the example, an resource returns a text payload derived from a URL parameter.
• Tools – callable actions that accept typed input (validated with Zod) and return a content payload. The tool demonstrates how an assistant can invoke server logic directly.
• Prompts – templates that generate conversational messages for the assistant. These enable dynamic user interactions without hard‑coding prompts in the client.

Because the server uses SSE, clients receive real‑time updates without polling. This is particularly valuable for long‑running tool executions or streaming content, allowing the assistant to provide incremental responses as data becomes available. The integration is seamless: any MCP‑compatible client (Claude Desktop, Cursor, etc.) can be pointed at the server’s SSE URL, and it will automatically discover the registered tools, resources, and prompts.

Typical use cases include:

• Internal tooling – expose internal APIs (e.g., database queries, file operations) to an assistant so developers can query data without leaving the chat interface.
• Custom prompt orchestration – create reusable prompt templates that adapt to user input, streamlining complex conversational flows.
• Real‑time data feeds – stream live metrics or logs to the assistant, enabling on‑the‑fly analysis and decision making.

Adonis MCP’s standout advantages are its tight integration with the AdonisJS ecosystem, type safety through Zod schemas, and zero‑configuration SSE transport. It gives developers a lightweight, opinionated way to turn any Adonis application into an AI‑ready backend without the overhead of building a separate MCP server from scratch.