We’ve been working on a new framework called MoroJS.
Why we built it:
Most existing frameworks (Express, Fastify, etc.) are excellent, but we found ourselves consistently trading off between three things:
•Strong type safety
•Edge/serverless deployment
•Features built for the AI era
MoroJS is an attempt to combine all three, while still pushing performance to the edge.
Key details:
•Performance: Benchmarked at ~68k synthetic req/sec and ~50k real-world req/sec
•TypeScript-first: Every route, param, and return strongly typed by default
•Edge/serverless ready: Runs on Node.js, Cloudflare Workers, AWS Lambda, Vercel
•AI-ready: Native support for building MCP (Model Context Protocol) servers, so APIs can plug directly into AI agents (e.g. Claude Desktop)
•Batteries included: Middleware ordering, Zod validation, caching, and rate limiting
Why we built it: Most existing frameworks (Express, Fastify, etc.) are excellent, but we found ourselves consistently trading off between three things: •Strong type safety •Edge/serverless deployment •Features built for the AI era
MoroJS is an attempt to combine all three, while still pushing performance to the edge.
Key details: •Performance: Benchmarked at ~68k synthetic req/sec and ~50k real-world req/sec •TypeScript-first: Every route, param, and return strongly typed by default •Edge/serverless ready: Runs on Node.js, Cloudflare Workers, AWS Lambda, Vercel •AI-ready: Native support for building MCP (Model Context Protocol) servers, so APIs can plug directly into AI agents (e.g. Claude Desktop) •Batteries included: Middleware ordering, Zod validation, caching, and rate limiting
Links: •Docs: morojs.com/docs •GitHub: github.com/Moro-JS