create-mcp-server — Scaffold an MCP Server in 30 Seconds
create-mcp-server is the `create-next-app` of MCP. `npm create mcp-server@latest` walks you through naming and template selection, then writes a complete TypeScript MCP server — package.json with the right SDK version pinned, tsconfig set up for stdio servers, src/index.ts with at least one working tool, README, and gitignore. Two templates ship in v0.1: `blank` (one example tool, your starting point for custom tools) and `markdown-rag` (exposes a ./content/ directory of markdown as MCP resources with a keyword search tool). Zero runtime dependencies in the CLI itself — pure Node + readline. Companion to the portfolio MCP server and the MCP Inspector demo, written to onboard other engineers into the MCP ecosystem.
Built in public to learn something or prove a pattern. Some are published packages, some are demos that live on this site.
- ▸Going from 'I want to build an MCP server' to a runnable hello-world is a 30-minute boilerplate exercise the first time
- ▸Each MCP client (Claude Desktop, Cursor, Cline, Zed) needs the same config snippet — generating it correctly is fiddly
- ▸Common patterns (RAG over a markdown dir, wrapping an API) recur, so each engineer reinvents the same scaffolding
- ▸Interactive Node + readline CLI with two starter templates: minimal (one tool) and markdown-RAG (resources + keyword search)
- ▸Auto-generated Claude Desktop config snippet with the absolute path pre-filled — paste it and you're done
- ▸Zero runtime dependencies in the CLI; templates pin @modelcontextprotocol/sdk and ship a working build pipeline (tsc → dist/)
Working MCP server scaffolded in under a minute, regardless of template chosen
Two production-shaped templates that onboard engineers into both 'blank canvas' and 'RAG over docs' patterns
Reusable scaffolder pattern: new templates require only a new ./templates/<name>/ directory
JobJam.io — AI Job Search & Application Platform
AI-powered job search platform: discover roles, evaluate fit, tailor applications, and close skill gaps. One-time pricing, no subscriptions.
Fathohm — Comprehension-Debt System of Record
See how much of your codebase no human understands. A treemap of your repo colored by human comprehension score, with a tracked, assignable metric for 'who understands what' in an AI-native codebase.
Grenz — Firewall for AI Agents
A local firewall for AI agents: scoped, revocable, observable permissions instead of your full GitHub token. Deny-by-default, MIT licensed, live at grenz.dev.