pmpt-mcp

pmpt-mcp

MCP server for pmptwiki — auto-saves project milestones with summaries, manages version history, and publishes AI-driven projects to pmptwiki.com community platform.

Category
访问服务器

README

<div align="center">

pmpt

Record and share your AI-driven product development journey.

npm version license node

Plan with 5 questions. Build with AI. Save every version. Share and reproduce.

Quick Start · Commands · MCP Server · Explore Projects

</div>


Demo

<img src="./demo.gif" alt="pmpt plan demo" width="600" />

Answer 5 questions → AI prompt auto-generated & copied to clipboard → Paste into Claude Code / Codex / Cursor → Start building!


Install

npm install -g pmpt-cli

Requires Node.js 18+


Quick Start

# 1. Install
npm i -g pmpt-cli

# 2. Initialize your project
pmpt init

# 3. Answer 5 questions → AI prompt auto-generated & copied
pmpt plan

# 4. Paste into Claude Code / Codex / Cursor → Build your product!

# 5. Save your progress anytime
pmpt save

# 6. Share with the community
pmpt login && pmpt publish

# Bonus: Explore what others are building
pmpt explore

Why pmpt?

Without pmpt With pmpt
Planning Stare at blank screen, write vague prompts Answer 5 guided questions, get structured AI prompt
Tracking Lose track of what you built and how Every version auto-saved with full history
Sharing Share finished code only Share the entire journey — others can reproduce it

The 5 Questions

pmpt plan asks just 5 questions to generate a complete AI prompt:

# Question Example
1 Project name my-budget-app
2 What to build A budget tracking app for freelancers
3 Additional context (optional) Simple UI, mobile-friendly
4 Key features Expense tracking; Monthly reports; CSV export
5 Tech stack (optional) React, Node.js — or let AI decide

The generated prompt is automatically copied to your clipboard. Just paste it into your favorite AI tool.


Commands

Local

Command Description
pmpt init Initialize project and start tracking
pmpt plan 5 questions → AI prompt (auto-copied to clipboard)
pmpt save Save current state as a snapshot
pmpt watch Auto-detect file changes and save versions
pmpt status Check project status, tracked files, and quality score
pmpt history View version history
pmpt diff v1 v2 Compare two versions (unified diff)
pmpt diff v3 Compare version vs working copy
pmpt squash v2 v5 Merge versions v2–v5 into one
pmpt export Export project as .pmpt file
pmpt import <file> Import from .pmpt file
pmpt recover Recover damaged pmpt.md via AI-generated prompt
pmpt mcp-setup Auto-configure MCP server for your AI tool
pmpt -v Show current CLI version

Platform

Command Description
pmpt login Authenticate via GitHub (one-time)
pmpt publish Publish your project (requires quality score ≥ 40)
pmpt update Quick re-publish: update content without changing metadata
pmpt edit Edit published project metadata (description, tags, category, visibility, related projects, product URL)
pmpt unpublish Remove a published project from pmptwiki
pmpt graduate Graduate a project — archive with badge, move to Hall of Fame
pmpt clone <slug> Clone and reproduce someone's project
pmpt explore Open pmptwiki.com/explore in your browser

Quality score below 40? pmpt copies an AI improvement prompt to your clipboard — paste it into your AI tool to get help improving your project.

See the full documentation at pmptwiki.com/docs


MCP Server

pmpt includes a built-in MCP server so AI tools can interact with pmpt directly. This means Claude Code, Cursor, and other MCP-compatible tools can save snapshots, check status, and review history without you typing commands.

Setup

pmpt-mcp is included when you install pmpt — no separate installation needed.

npm install -g pmpt-cli   # pmpt + pmpt-mcp both installed

Automatic Setup (Recommended)

pmpt mcp-setup

Auto-detects the pmpt-mcp binary path, lets you choose your AI tool (Claude Code, Cursor, or .mcp.json), and writes the config. Solves PATH issues common with nvm.

Manual Setup

Add to your .mcp.json (or IDE MCP config):

{
  "mcpServers": {
    "pmpt": {
      "command": "pmpt-mcp"
    }
  }
}

nvm users: If your AI tool can't find pmpt-mcp, use pmpt mcp-setup or replace "pmpt-mcp" with the absolute path (run which pmpt-mcp to find it).

Available Tools

Tool Description
pmpt_plan_questions Get planning questions to ask the user conversationally
pmpt_plan Submit answers to generate AI prompt and project docs
pmpt_save Save a snapshot with detailed summary
pmpt_update_doc Check off features, add progress notes, backfill summaries
pmpt_log_decision Record architectural/technical decisions with reasoning
pmpt_read_context Read full project context (plan, docs, history, quality)
pmpt_status Check tracked files, snapshot count, and quality score
pmpt_history View version history with git commit info
pmpt_diff Compare two versions, or a version against working copy
pmpt_edit_plan Edit plan fields — regenerates plan.md and pmpt.ai.md
pmpt_quality Check quality score and publish readiness
pmpt_publish Publish project to pmptwiki.com

All tools accept an optional projectPath parameter (defaults to cwd).


How It Works

  You                          pmpt                         AI Tool
   │                            │                              │
   ├── pmpt init ──────────────→│                              │
   │                            │  Creates .pmpt/ project      │
   │                            │                              │
   ├── pmpt plan ──────────────→│                              │
   │   (answer 5 questions)     │  Generates AI prompt         │
   │                            │  → Copied to clipboard       │
   │                            │                              │
   ├── Ctrl+V ─────────────────────────────────────────────────→│
   │                            │                    Builds your product
   │                            │                              │
   ├── pmpt save ──────────────→│                              │
   │                            │  Snapshots version history   │
   │                            │                              │
   ├── pmpt publish ───────────→│                              │
   │                            │  Shares with community       │
   │                            │                              │
   └── Others: pmpt clone ─────→│  Reproduces your journey     │

Project Structure

your-project/
└── .pmpt/
    ├── config.json        # Project configuration
    ├── docs/              # Generated documents
    │   ├── plan.md        # Product plan (features checklist)
    │   ├── pmpt.md        # Progress tracking & decisions
    │   └── pmpt.ai.md     # AI-ready prompt (project context & instructions)
    └── .history/          # Auto-saved version history
        ├── v1-2024-02-20/
        ├── v2-2024-02-21/
        └── ...

.pmpt File Format

A single portable file containing your entire development journey:

{
  "schemaVersion": "1.0",
  "meta": {
    "projectName": "my-budget-app",
    "description": "Budget tracking app for freelancers",
    "createdAt": "2024-02-20T10:00:00Z"
  },
  "plan": {
    "productIdea": "A budget tracking app...",
    "coreFeatures": "Expense tracking; Monthly reports",
    "techStack": "React, Node.js"
  },
  "docs": {
    "plan.md": "...",
    "pmpt.md": "...",
    "pmpt.ai.md": "..."
  },
  "history": [
    { "version": 1, "timestamp": "...", "files": {} },
    { "version": 2, "timestamp": "...", "files": {} }
  ]
}

Share it. Clone it. Reproduce it.


Use Cases

  • Have an idea but no coding skills? — Answer 5 questions, paste the prompt into AI, and start building
  • Startup founders — Document your MVP creation process from day one
  • Content creators — Share your AI-assisted building journey as reproducible content
  • Learners — Clone published projects to study how others build with AI

Explore Projects

Don't know what to build? See what others have created with AI.

# Open the explore page
pmpt explore

# Found something interesting? Clone it and make it yours
pmpt clone budget-tracker-app

Explore Projects on pmptwiki.com →

Clone any project to see how it was planned, what prompts were used, and how it evolved step by step. The clone output shows the product idea, tech stack, and full version history.


Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request.


Links


License

MIT


<div align="center">

If pmpt helps you build something, give it a ⭐

Explore what others are building →

Buy Me A Coffee

</div>

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选