JSPM

spec2code

3.1.10
  • ESM via JSPM
  • ES Module Entrypoint
  • Export Map
  • Keywords
  • License
  • Repository URL
  • TypeScript Types
  • README
  • Created
  • Published
  • Downloads 4
  • Score
    100M100P100Q30118F
  • License MIT

Specification-Driven Software Development with LLM-Powered Automation

Package Exports

  • spec2code
  • spec2code/dist/index.js

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (spec2code) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Spec2Code: Turn Ideas Into Code Automatically 🚀

npm version License: MIT TypeScript

Transform any idea into a complete, working application using AI in Cursor - works with ANY LLM (Claude, GPT-4, Gemini, etc.).

Note: The tool will be open-sourced in the future. Github link is private for now.

🚀 Quick Start

Installation

npm install -g spec2code

Add to Cursor

Add the following configuration to your Cursor mcp.json file:

{
  "mcpServers": {
    "spec2code": {
      "command": "spec2code",
      "args": [],
      "env": {
        "PROJECT_ROOT": "${workspaceFolder}"
      }
    }
  }
}

Basic Usage

# 1. Generate specification
/sdd_specify "Build a crowdfunding platform with Next.js, TypeScript, and Tailwind"

# 2. Create implementation plan
/sdd_plan

# 3. Generate tasks (33 tasks across 4 phases)
/sdd_tasks

# 4. Execute phase-by-phase
/sdd_implement phase=1  # Setup & Foundations
/sdd_implement phase=2  # Core Implementation
/sdd_implement phase=3  # UI Development
/sdd_implement phase=4  # Testing & Deployment

🛠️ Commands

  • /sdd_specify - Generate project specifications (creates specs/spec.md)
  • /sdd_plan - Create implementation plan (creates specs/plan.md)
  • /sdd_tasks - Generate 33 tasks across 4 phases (creates specs/phaseX-tasks.md)
  • /sdd_implement phase=X - Execute phase-by-phase implementation
  • /sdd_implement phase=X task=Y - Execute a specific task
  • /sdd_workflow - Run complete automated workflow (specify → plan → tasks → implement)

🌟 Key Features

  • Complete Task Execution - AI completes all tasks fully, including verification, before proceeding to next task
  • AI-Driven TODO Creation - Each task generates its own TODO list by analyzing requirements
  • Platform Detection - Automatically adapts to web, mobile, desktop, backend, and AI platforms
  • Architecture-Aware - Detects BaaS (Firebase/Supabase), Traditional Backend, Serverless, and Hybrid patterns, adapting tasks accordingly
  • Task Verification - All tasks must be individually verified with proof (terminal output, code snippets) before completion
  • Time Estimation - Use Case Points (UCP) and PERT-based estimates with AI speedup factors (50x faster)
  • Production-Ready - Complete test suites, error handling, documentation, and deployment

🏗️ How It Works

Workflow:

  1. /sdd_specify → Creates specs/spec.md with functional requirements
  2. /sdd_plan → Creates specs/plan.md with implementation strategy
  3. /sdd_tasks → Creates specs/phase1-tasks.md through phase4-tasks.md (33 tasks total)
  4. /sdd_implement phase=X → Executes tasks sequentially with verification

Files Generated:

  • specs/spec.md - Project specification
  • specs/plan.md - Implementation plan
  • specs/phase1-tasks.md to phase4-tasks.md - Task breakdowns (9 + 8 + 9 + 7 = 33 tasks)
  • specs/phaseX-status.json - Task completion tracking

Platforms Supported:

  • Frontend: Web (Next.js, React), Mobile (React Native, Flutter), Desktop (Electron, Tauri)
  • Backend: Node.js, Python, Java, .NET
  • BaaS: Firebase, Supabase, AWS Amplify
  • Serverless: AWS Lambda, Google Cloud Functions, Vercel Functions
  • AI: TensorFlow, PyTorch

🏆 Real-World Examples

See Complete applications built with this tool:

  • Recipe Finder (3 hours)
  • Invoice Generator (5.5 hours)
  • Kanban Board (5.5 hours)
  • Video Conferencing (6.5 hours)
  • Learning Management System (5 hours)
  • AR Home Decorator (8 hours)

Note: Some apps may have database connection issues due to Supabase auto-deletion.

📋 Requirements

  • Node.js 18+
  • Cursor IDE (with MCP enabled)
  • Any LLM (Claude, GPT-4, Gemini, etc.)

👥 Perfect For

  • Beginners: Turn ideas into apps without coding knowledge
  • Developers: Speed up development with AI assistance
  • Startups: Rapidly prototype and build MVPs
  • Students: Learn by seeing complete implementations
  • Teams: Standardize development processes

Transform your development workflow with AI-powered specification-driven development! 🎉