[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-langgptai--LangGPT":3,"tool-langgptai--LangGPT":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":80,"owner_twitter":79,"owner_website":79,"owner_url":81,"languages":82,"stars":99,"forks":100,"last_commit_at":101,"license":102,"difficulty_score":103,"env_os":104,"env_gpu":104,"env_ram":104,"env_deps":105,"category_tags":108,"github_topics":109,"view_count":23,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":124,"updated_at":125,"faqs":126,"releases":162},2323,"langgptai\u002FLangGPT","LangGPT","LangGPT: Empowering everyone to become a prompt expert! 🚀  📌 结构化提示词（Structured Prompt）提出者 📌 元提示词（Meta-Prompt）发起者   📌 最流行的提示词落地范式 | Language of GPT  The pioneering framework for structured & meta-prompt design 10,000+ ⭐ | Battle-tested by thousands of users worldwide  Created by 云中江树","LangGPT 是一套专为大语言模型设计的结构化提示词框架，旨在让每个人都能轻松成为提示词专家。它将编写提示词的过程类比为编程，提供了一套系统化的模板和方法论，帮助用户快速构建高质量、可复用的指令。\n\n传统提示词工程往往依赖零散的经验积累和反复试错，效率较低且难以标准化。LangGPT 通过引入层级分明的结构（如角色设定、技能描述、规则约束、工作流等），将这一过程转化为模块化、逻辑清晰的“提示词语言”。这不仅大幅降低了创作门槛，还确保了输出结果的稳定性和一致性。\n\n无论是希望提升工作效率的普通用户、需要精准控制模型输出的开发者，还是从事相关研究的研究人员，都能从 LangGPT 中受益。其独特的技术亮点在于提出了“元提示词”概念，并拥有经过全球数万名用户验证的成熟生态，甚至获得了学术论文的支持。用户既可以直接使用官方提供的自动化工具生成提示词，也可以基于丰富的示例库手动定制，更支持在 Claude Code 等环境中以技能插件形式调用，灵活适配各种使用场景。","# 🚀 LangGPT — Empowering Everyone to Create High-Quality Prompts!\n\n\u003Cdiv align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_af068c4c9b9d.png\" width=\"60%\" height=\"auto\">\n\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-blue.svg)](\u002FLICENSE)\n[![Status](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fstatus-active-success.svg)]()\n[![Paper](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2402.16929-b31b1b.svg)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.16929)\n[![Stars](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_7378c5787671.png)](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT)\n\n[English](README.md) | [简体中文](README_zh.md) | [日本語](README_ja.md)\n\n[Quick Start](#-quick-start) | [Theoretical Foundations](#-theoretical-foundations) | [Ecosystem](#-langgpt-ecosystem) | [Community](http:\u002F\u002Ffeishu.langgpt.ai)\n\n\u003C\u002Fdiv>\n\n---\n\n## 📖 What is LangGPT?\n\n**LangGPT is a structured, reusable prompt design framework** that enables anyone to create high-quality prompts for Large Language Models. Think of it as a **\"programming language for prompts\"** — systematic, template-based, and infinitely scalable.\n\n### Why LangGPT?\n\nTraditional prompt engineering relies on scattered tips and trial-and-error. LangGPT transforms this chaos into a structured methodology:\n\n- 🎯 **Structured Templates** — Hierarchical organization inspired by programming paradigms\n- 🔄 **Reusability** — Create once, adapt infinitely like code modules  \n- 📦 **Modularity** — Variables, commands, and conditional logic at your fingertips\n- ⚡ **Efficiency** — Go from idea to working prompt in minutes\n- 🌍 **Community-Driven** — 11,000+ stars, battle-tested by thousands of users\n\n> **Academic Foundation**: Published research at [arXiv:2402.16929](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.16929) | [中文版](Papers\u002FLangGPT_paper_cn.md)\n\n---\n\n## 🚀 Quick Start\n\n### Method 1: Use Automated Tools (Fastest)\n\nLet AI create prompts for you:\n\n- **[LangGPT GPTs](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt)** — Full-featured generator (GPT-4)\n- **[Kimi+ LangGPT](https:\u002F\u002Fkimi.moonshot.cn\u002Fkimiplus\u002Fconpg00t7lagbbsfqkq0)** — For Moonshot Kimi users\n- **[PromptGPT](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YKe3gmydD-promptgpt)** — Lite version (GPT-3.5)\n\n### Method 2: Master the Template (5 Minutes)\n\nBasic LangGPT structure:\n\n```markdown\n# Role: Your_Role_Name\n\n## Profile\n- Author: YourName\n- Version: 1.0\n- Language: English\n- Description: Clear role description and core capabilities\n\n## Goal\n- Outcome: What concrete result\u002Foutcome should be delivered for the user\u002Fsession\n- Done Criteria: Clear acceptance criteria (how we know it’s finished and good)\n- Non-Goals: What is explicitly out of scope to avoid scope creep\n\n### Skill-1\n1. Specific skill description\n2. Expected behavior and output\n\n## Rules\n1. Don't break character under any circumstance\n2. Don't make up facts or hallucinate\n\n## Workflow\n1. Analyze user input and identify intent\n2. Apply relevant skills systematically\n3. Deliver structured, actionable output\n\n## Initialization\nAs a\u002Fan \u003CRole>, you must follow the \u003CRules>, you must talk to user in default \u003CLanguage>, you must greet the user. Then introduce yourself and introduce the \u003CWorkflow>.\n```\n\n**Prerequisites**: Basic Markdown knowledge ([Quick Guide](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fget-started\u002Fwriting-on-github\u002Fgetting-started-with-writing-and-formatting-on-github\u002Fbasic-writing-and-formatting-syntax)) | GPT-4 or Claude recommended\n\n### Method 3: Start from Examples\n\nExplore our [example library](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe) and adapt proven templates to your needs.\n\n### Method 4: Claude Code Skill (Recommended)\n\nIf you use [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code), install the LangGPT Skill to get structured prompt writing capabilities:\n\n**Installation:**\n\n1. Download [langgpt.skill](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Freleases)\n2. Extract to `~\u002F.claude\u002Fskills\u002F` directory\n3. Type `\u002Flanggpt` in Claude Code to use\n\n**Skill Features:**\n- 📝 Structured prompt templates (Role, Profile, Skills, Rules, Workflow)\n- 📚 Rich example library (FitnessGPT, Poet, Xiaohongshu Master, Name Master, etc.)\n- 🔧 Advanced techniques: variables, commands, conditional logic\n- 🎯 Model compatibility guide (GPT-4, Claude, GPT-3.5)\n\n---\n\n## 🧠 Theoretical Foundations\n\nBefore diving into tactics, understand the principles. These essays explore the philosophy behind effective prompting:\n\n- **[对话动力学](Docs\u002F对话动力学.md)** — The dynamics of human-AI dialogue\n- **[五种理性](Docs\u002F五种理性.md)** — Five types of rationality in prompt design\n- **[镜像性倾向](Docs\u002F镜像性倾向.md)** — Mirror tendencies in LLM behavior\n- **[统计重力井和边缘表达](Docs\u002F统计重力井和边缘表达.md)** — Statistical gravity well and edge expression\n- **[关系表达](Docs\u002F关系表达.md)** — Expressing relationships in prompts\n- **[看见与言说](Docs\u002F看见与言说.md)** — Seeing and articulation in AI interaction  \n- **[Prompt 的本质](Docs\u002FPrompt的本质.md)** — The essence and nature of prompts\n- **[面向结果的提示词写作方法](Docs\u002F面向结果的提示词写作方法.md)** — Writing prompts that focus on achieving desired outcomes\n- **[AI意识](Docs\u002FAI意识.md)** — Understanding the role of AI in human-AI interaction\n- **[AI时代的新管理：机器负责优化，人类定义应该](Docs\u002FAI时代的新管理：机器负责优化，人类定义应该.md)** — The new management in the AI era: machines optimize, humans define the criteria\n\n*These foundational insights will transform how you think about prompts.*\n\n---\n\n## 💡 Core Concepts\n\n### 1. Structured Roles\n\nDefine AI personas through clear, modular sections:\n\n| Section | Purpose | Example |\n|---------|---------|---------|\n| **Role** | Role name\u002Ftitle | \"逻辑学家\" \u002F \"Expert Analyst\" \u002F \"FitnessGPT\" |\n| **Profile** | Identity and capabilities | \"Expert Python developer with 10 years experience\" |\n| **Goal**           | Desired outcome, done criteria, and non-goals for this session\u002Ftask | “Refactor a prompt into a reusable template; acceptance criteria: pass three structured checks; non-goal: rewriting the business logic.”                    |\n| **Skills** | Specific abilities | \"Debug complex code, optimize performance\" |\n| **Rules** | Boundaries and constraints | \"Never execute destructive commands\" |\n| **Workflow** | Interaction logic | \"1. Analyze → 2. Plan → 3. Execute\" |\n| **Initialization** | Opening message and setup | \"As a \u003CRole>, I will greet you and introduce the \u003CWorkflow>\" |\n\n### 2. Variables and References\n\nUse `\u003CVariable>` syntax for dynamic content:\n\n```markdown\nAs a \u003CRole>, you must follow \u003CRules> and communicate in \u003CLanguage>\n```\n\nThis creates self-referential prompts that maintain consistency across complex instructions.\n\n### 3. Commands\n\nDefine reusable actions for better UX:\n\n```markdown\n## Commands\n- Prefix: \"\u002F\"\n- Commands:\n    - help: Display all available commands\n    - continue: Resume interrupted output\n    - improve: Enhance current response with deeper analysis\n```\n\n### 4. Conditional Logic\n\nAdd intelligence to your prompts:\n\n```markdown\nIf user provides [code], then analyze and suggest improvements\nElse if user asks [question], then provide detailed explanation\nElse, prompt for clarification\n```\n\n### 5. Advanced Techniques\n\n**Reminders** — Combat context loss in long conversations:\n```markdown\n## Reminder\n1. Always check role settings before responding\n2. Current language: \u003CLanguage>, Active rules: \u003CRules>\n```\n\n**Alternative Formats** — Use JSON\u002FYAML when markdown isn't ideal:\n```yaml\nrole: DataAnalyst\nprofile:\n  version: \"2.0\"\n  language: \"Python\"\nskills:\n  - statistical_analysis\n  - data_visualization\n```\n\n---\n\n## 🌟 Featured Examples\n\n| Prompt | Description | Link |\n|--------|-------------|------|\n| 🎯 **FitnessGPT** | Personalized diet and workout planner | [View](examples\u002FFitnessGPT.md) |\n| 💻 **Code Master CAN** | Advanced coding assistant with debugging expertise | [View](examples\u002Fcode_anything_now\u002FChatGPT-Code_Anything_Now_en.md) |\n| ✍️ **Xiaohongshu Writer** | Viral social media content generator | [View](examples\u002Fchinese_xiaohongshu_writer\u002F) |\n| 🎨 **Chinese Poet** | Classical poetry composer in traditional styles | [View](examples\u002Fchinese_poet\u002F) |\n\n[Browse 100+ more examples →](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe)\n\n---\n\n## 📚 Learning Resources\n\n### Essential Guides\n\n| Resource | Description | Date |\n|----------|-------------|------|\n| [Academic Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.16929) | LangGPT: Rethinking Structured Reusable Prompt Design ([中文](Papers\u002FLangGPT_paper_cn.md)) | Feb 2024 |\n| [Structured Prompts Guide](Docs\u002FHowToWritestructuredPrompts.md) | Comprehensive tutorial on building high-performance prompts | Jul 2023 |\n| [Prompt Chains](Docs\u002FPromptChain.md) | Multi-prompt collaboration and task decomposition strategies | Aug 2023 |\n| [Video Tutorial](https:\u002F\u002Fwww.bilibili.com\u002Fvideo\u002FBV1rj411q78a) | BiliBili walkthrough (by AIGCLINK) | Sep 2023 |\n\n### Advanced Topics\n\n- **[推理模型提示方法变革](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FFLY0sy1jYv6eT9151Yz_jw)** — Paradigm shift from procedural to goal-oriented prompting\n- **[提示词的道和术](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FAYMWwBPaSih46WkAo9jcfKkfntg)** — Philosophy and practice of prompt engineering by 李继刚\n- **[企业级提示词工程](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FUTyswvusTiRw0TkZLI5cIG0Tnhc)** — Building production-ready prompt systems (百川智能)\n- **[多模态提示词](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FAan9NXO_vEZ9h0YrugpoGQ)** — GPT-4V and multi-modal prompting techniques\n- **[提示词攻击与防护](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FaaABXnxRqDF716qRk79wYQ)** — Security: prompt injection, jailbreaks, and defenses\n- **[大模型绘画指南](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FbJbZ9bwPXxlpyREqLKhDvA)** — AI image generation with structured prompts\n\n### Community Hub\n\n**[Feishu Knowledge Base](http:\u002F\u002Ffeishu.langgpt.ai)** — Curated resources, templates, and community contributions\n\n---\n\n## 🎨 LangGPT Ecosystem\n\n### Core Framework & Tools\n\n| Project | Description | Stars |\n|---------|-------------|-------|\n| **[LangGPT](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT)** | Core framework and methodology | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_7378c5787671.png) |\n| **[PromptVer](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FPromptVer)** | Semantic versioning for prompts — version control like Git | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_bb0f9a9c5a24.png) |\n| **[PromptShow](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FPromptShow)** | Create beautiful prompt images ([Try it](https:\u002F\u002Fshow.langgpt.ai\u002F)) | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_1bb2967783e0.png) |\n| **[Minstrel](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FMinstrel)** | Multi-agent system for auto-generating prompts | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_eba16dc6c6e5.png) |\n\n### Model-Specific Prompt Collections\n\nRather than writing prompts as procedures, write the persona. Writing prompts as procedures gives the model steps and tools. Writing prompts as a persona gives the model a worldview, motivations, a value system, and a preference profile. Below are prompts that Yunzhong Jiangshu wrote while studying some well-known figures. \n\n* [巴菲特AI分身](Prompts\u002F巴菲特AI分身.md)\n* [梵高AI分身](Prompts\u002F梵高AI分身.md)\n* [马斯克AI分身](Prompts\u002F马斯克AI分身.md)\n* [段永平AI分身](Prompts\u002F段永平AI分身.md)\n\nCurated, optimized prompts for different AI models:\n\n| Collection | Target Model | Stars |\n|------------|--------------|-------|\n| [wonderful-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fwonderful-prompts) | ChatGPT (Chinese) | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_892f31557870.png) |\n| [awesome-claude-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-claude-prompts) | Anthropic Claude | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_209a3e7267bf.png) |\n| [awesome-deepseek-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-deepseek-prompts) | DeepSeek & R1 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_527bca0e3e0e.png) |\n| [awesome-gemini-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-gemini-prompts) | Google Gemini | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_b6aafed0d0c6.png) |\n| [awesome-grok-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-grok-prompts) | xAI Grok | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_250d00337478.png) |\n| [qwen-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fqwen-prompts) | Alibaba Qwen | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_f6b05c435961.png) |\n| [awesome-llama-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-llama-prompts) | Meta Llama 2\u002F3 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_eac74f96b77c.png) |\n| [awesome-doubao-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-doubao-prompts) | ByteDance Doubao | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_18a2a27fd34a.png) |\n| [awesome-system-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-system-prompts) | System prompts from AI tools | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_74c979c3d970.png) |\n\n### Specialized Domains\n\n| Repository | Focus Area | Stars |\n|------------|------------|-------|\n| [Awesome-Multimodal-Prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FAwesome-Multimodal-Prompts) | GPT-4V, DALL-E 3, image\u002Fvideo prompts | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_14f2d9dfa793.png) |\n| [deep-research-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fdeep-research-prompts) | Deep research across models | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_ae3a202e5458.png) |\n| [awesome-voice-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-voice-prompts) | Voice AI and conversational agents | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_4e7a0e12702a.png) |\n| [GraphRAG-Prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FGraphRAG-Prompts) | Graph-based retrieval prompts | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_d1168ca9fa14.png) |\n| [LLM-Jailbreaks](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLLM-Jailbreaks) | Security research and defenses | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_204f9389e2b1.png) |\n\n### Applications\n\n| Project | Description | Stars |\n|---------|-------------|-------|\n| [BookAI](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FBookAI) | AI-powered book generation | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_4e178c446c27.png) |\n| [AI-Resume](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FAI-Resume) | Beautiful resumes with Claude Artifacts | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_e2205012b76e.png) |\n\n---\n\n## 🛠️ Popular GPTs Built with LangGPT\n\nTransform ChatGPT with these specialized assistants:\n\n| GPT | Purpose | Link |\n|-----|---------|------|\n| 🎯 **LangGPT Expert** | Auto-generate structured prompts | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt) |\n| ✍️ **PromptGPT** | Professional prompt engineer | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YKe3gmydD-promptgpt) |\n| 🧠 **SmartGPT-5** | Never lazy, always diligent assistant | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-sRQtxpN4C-smartgpt-5) |\n| 💻 **Coding Expert** | Comprehensive programming assistant | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-ky06YjwaP-coding-expert) |\n| 📊 **Data Table GPT** | Transform messy data into clean tables | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-nb6RjxHsb-data-table-gpt) |\n| 🔥 **PytorchGPT** | PyTorch code specialist | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-kyj3zKyHK-pytorchgpt) |\n| 🎨 **LogoGPT** | Professional logo designer | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-wdz2JlUBv-logogpt) |\n| 📄 **PDF Reader** | Deep document analysis and extraction | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YaMjCVW0t-pdf-reader) |\n| 🏅 **MathGPT** | Precise mathematical problem solver | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-UIOlPhTjK-mathgpt) |\n| 📝 **WriteGPT** | Professional writing across industries | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-jwTMtRiL8-writegpt) |\n| 🎙️ **时事热评员** | Current events commentator | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-gbfs6fy7c-shi-shi-re-ping-yuan) |\n| 🎀 **翻译大小姐** | Elegant Chinese translations | [Launch](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-2V90YGvVD-fan-yi-da-xiao-jie) |\n\n[Discover 20+ more GPTs →](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT#langgpt-gpts)\n\n---\n\n## 🤝 Contributing\n\nWe welcome all contributions to make LangGPT better!\n\n### How You Can Help\n\n1. ⭐ **Star and share** — Increase visibility and help others discover LangGPT\n2. 📝 **Submit examples** — Share your successful prompts built with LangGPT\n3. 🆕 **Propose templates** — Create new templates beyond the Role structure\n4. 📖 **Improve docs** — Fix typos, clarify instructions, add translations\n5. 💡 **Suggest features** — Open issues with ideas for new capabilities\n6. 🔧 **Code contributions** — Help build tools, utilities, and integrations\n\n### Getting Started\n\nNew to GitHub contributions? Check out this [GitHub Minimal Contribution Guide](https:\u002F\u002Fgithub.com\u002Fdatawhalechina\u002FDOPMC\u002Fblob\u002Fmain\u002FGITHUB.md)\n\n---\n\n## 📊 Star History\n\n[![Star History Chart](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_43054fb116f5.png)](https:\u002F\u002Fstar-history.com\u002F#langgptai\u002FLangGPT&Date)\n\n---\n\n## 📄 Citation\n\nIf you use LangGPT in research or projects, please cite:\n\n```bibtex\n@misc{wang2024langgpt,\n      title={LangGPT: Rethinking Structured Reusable Prompt Design Framework for LLMs from the Programming Language}, \n      author={Ming Wang and Yuanzhong Liu and Xiaoming Zhang and Songlian Li and Yijie Huang and Chi Zhang and Daling Wang and Shi Feng and Jigang Li},\n      year={2024},\n      eprint={2402.16929},\n      archivePrefix={arXiv},\n      primaryClass={cs.SE}\n}\n```\n\n---\n\n## 🙏 Acknowledgments\n\nLangGPT was inspired by excellent projects:\n\n- [Mr.-Ranedeer-AI-Tutor](https:\u002F\u002Fgithub.com\u002FJushBJJ\u002FMr.-Ranedeer-AI-Tutor) — Structured tutoring prompts\n- [Auto-GPT](https:\u002F\u002Fgithub.com\u002FSignificant-Gravitas\u002FAuto-GPT) — Autonomous AI agents\n- [SoM](https:\u002F\u002Fgithub.com\u002FSkalskiP\u002FSoM) — Set of Mark prompting\n- [yolov10](https:\u002F\u002Fgithub.com\u002FTHU-MIG\u002Fyolov10) — Computer vision innovations\n\n### Projects Built with LangGPT\n\nWe're proud to see LangGPT principles applied in the wild:\n- **[Prompt Optimizer](https:\u002F\u002Fgithub.com\u002Flinshenkx\u002Fprompt-optimizer)** — Intelligent prompt optimization tool leveraging LangGPT methodology\n- **[securityGPT](https:\u002F\u002Fgithub.com\u002Frryuliu\u002FsecurityGPT)** — Secure prompt protection against leaks\n- **[AIPainting-Structured-Prompts](https:\u002F\u002Fgithub.com\u002Fzhutyler21\u002FAIPainting-Structured-Prompts)** — Structured prompts for AI art generation\n\n---\n\n## 📬 Connect With Us\n\n### Author\n\n**云中江树 (Yun Zhong Jiang Shu)**\n- 📱 WeChat Official Account: **「云中江树」**\n- 💼 Creator of LangGPT Framework\n- 🎓 Prompt Engineering Researcher\n\n### Community\n\n- 📚 [Knowledge Base](http:\u002F\u002Ffeishu.langgpt.ai) — Comprehensive documentation\n- 🐦 [Twitter\u002FX](https:\u002F\u002Ftwitter.com\u002Flanggptai) — Latest updates\n- 💬 [GitHub Discussions](https:\u002F\u002Fgithub.com\u002Flanggptai) — Community forum\n- 📧 Email: contact@langgpt.ai\n\n---\n\n\u003Cdiv align=\"center\">\n\n**[⬆ Back to Top](#-langgpt--empowering-everyone-to-create-high-quality-prompts)**\n\nMade with ❤️ by the [langgptai Community](https:\u002F\u002Fgithub.com\u002Flanggptai)\n\n*Empowering everyone to become a prompt expert* 🚀\n\n\u003C\u002Fdiv>","# 🚀 LangGPT — 让每个人都能创建高质量提示词！\n\n\u003Cdiv align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_af068c4c9b9d.png\" width=\"60%\" height=\"auto\">\n\n[![许可证](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-blue.svg)](\u002FLICENSE)\n[![状态](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fstatus-active-success.svg)]()\n[![论文](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2402.16929-b31b1b.svg)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.16929)\n[![星标](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_7378c5787671.png)](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT)\n\n[English](README.md) | [简体中文](README_zh.md) | [日本語](README_ja.md)\n\n[快速入门](#-quick-start) | [理论基础](#-theoretical-foundations) | [生态系统](#-langgpt-ecosystem) | [社区](http:\u002F\u002Ffeishu.langgpt.ai)\n\n\u003C\u002Fdiv>\n\n---\n\n## 📖 什么是LangGPT？\n\n**LangGPT是一个结构化、可复用的提示词设计框架**，使任何人都能为大型语言模型创建高质量的提示词。你可以把它想象成一种“提示词编程语言”——系统化、基于模板且可无限扩展。\n\n### 为什么选择LangGPT？\n\n传统的提示词工程依赖于零散的技巧和反复试验。而LangGPT则将这种混乱转化为一种结构化的方法论：\n\n- 🎯 **结构化模板** — 受编程范式启发的层次化组织\n- 🔄 **可复用性** — 一次创建，像代码模块一样无限适配  \n- 📦 **模块化** — 变量、命令和条件逻辑触手可及\n- ⚡ **效率** — 几分钟内即可从想法生成可用的提示词\n- 🌍 **社区驱动** — 超过11,000颗星，经数千用户实战检验\n\n> **学术基础**：发表于[arXiv:2402.16929](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.16929)的研究 | [中文版](Papers\u002FLangGPT_paper_cn.md)\n\n---\n\n## 🚀 快速入门\n\n### 方法1：使用自动化工具（最快）\n\n让AI为你自动生成提示词：\n\n- **[LangGPT GPTs](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt)** — 全功能生成器（GPT-4）\n- **[Kimi+ LangGPT](https:\u002F\u002Fkimi.moonshot.cn\u002Fkimiplus\u002Fconpg00t7lagbbsfqkq0)** — 针对Moonshot Kimi用户\n- **[PromptGPT](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YKe3gmydD-promptgpt)** — 精简版（GPT-3.5）\n\n### 方法2：掌握模板（5分钟）\n\n基本LangGPT结构如下：\n\n```markdown\n# 角色: 您的角色名称\n\n## 个人简介\n- 作者: 您的名字\n- 版本: 1.0\n- 语言: 英语\n- 描述: 清晰的角色描述及核心能力\n\n## 目标\n- 结果: 用户或本次会话应获得的具体成果\n- 完成标准: 明确的验收标准（如何判断任务已完成且质量合格）\n- 非目标: 明确不在范围内的内容，以避免范围蔓延\n\n### 技能-1\n1. 具体技能描述\n2. 预期行为与输出\n\n## 规则\n1. 任何情况下都不得脱离角色设定\n2. 不得编造事实或产生幻觉\n\n## 工作流程\n1. 分析用户输入并识别意图\n2. 系统性地应用相关技能\n3. 提供结构化、可操作的输出\n\n## 初始化\n作为\u003C角色>,您必须遵守\u003C规则>,必须以默认\u003C语言>与用户交流，并向用户问好。随后介绍自己及\u003C工作流程>。\n```\n\n**前提条件**：具备基础Markdown知识（[快速指南](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fget-started\u002Fwriting-on-github\u002Fgetting-started-with-writing-and-formatting-on-github\u002Fbasic-writing-and-formatting-syntax)) | 推荐使用GPT-4或Claude\n\n### 方法3：从示例入手\n\n浏览我们的[示例库](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe)，根据需求调整经过验证的模板。\n\n### 方法4：Claude Code技能（推荐）\n\n如果你使用[Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code)，可以安装LangGPT技能，获得结构化的提示词编写能力：\n\n**安装步骤：**\n\n1. 下载[langgpt.skill](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Freleases)\n2. 解压到`~\u002F.claude\u002Fskills\u002F`目录\n3. 在Claude Code中输入`\u002Flanggpt`即可使用\n\n**技能特点：**\n- 📝 结构化提示词模板（角色、简介、技能、规则、流程）\n- 📚 丰富的示例库（FitnessGPT、诗人、小红书达人、取名大师等）\n- 🔧 高级技巧：变量、命令、条件逻辑\n- 🎯 模型兼容性指南（GPT-4、Claude、GPT-3.5）\n\n---\n\n## 🧠 理论基础\n\n在深入实践之前，先了解其背后的原理。这些文章探讨了有效提示词设计的哲学思想：\n\n- **[对话动力学](Docs\u002F对话动力学.md)** — 人机对话的动力机制\n- **[五种理性](Docs\u002F五种理性.md)** — 提示词设计中的五种理性类型\n- **[镜像性倾向](Docs\u002F镜像性倾向.md)** — 大型语言模型行为中的镜像倾向\n- **[统计重力井和边缘表达](Docs\u002F统计重力井和边缘表达.md)** — 统计重力井与边缘表达\n- **[关系表达](Docs\u002F关系表达.md)** — 在提示词中表达关系\n- **[看见与言说](Docs\u002F看见与言说.md)** — AI交互中的感知与表达\n- **[提示词的本质](Docs\u002FPrompt的本质.md)** — 提示词的本质与特性\n- **[面向结果的提示词写作方法](Docs\u002F面向结果的提示词写作方法.md)** — 以实现预期结果为导向的提示词写作\n- **[AI意识](Docs\u002FAI意识.md)** — 理解AI在人机交互中的角色\n- **[AI时代的新型管理：机器负责优化，人类定义应该](Docs\u002FAI时代的新管理：机器负责优化，人类定义应该.md)** — AI时代的新管理模式：机器负责优化，人类定义标准\n\n*这些基础洞见将彻底改变你对提示词的看法。*\n\n---\n\n## 💡 核心概念\n\n### 1. 结构化角色\n\n通过清晰、模块化的部分定义AI人格：\n\n| 部分       | 作用                     | 示例                           |\n|------------|--------------------------|--------------------------------|\n| **角色**   | 角色名称\u002F头衔            | “逻辑学家” \u002F “专家分析师” \u002F “FitnessGPT” |\n| **简介**   | 身份与能力               | “拥有10年经验的Python专家开发人员” |\n| **目标**   | 本次会话\u002F任务的期望成果、完成标准及非目标 | “将提示词重构为可复用模板；验收标准：通过三次结构化检查；非目标：重写业务逻辑。” |\n| **技能**   | 具体能力                 | “调试复杂代码，优化性能”         |\n| **规则**   | 边界与约束               | “绝不执行破坏性命令”             |\n| **流程**   | 交互逻辑                 | “1. 分析 → 2. 计划 → 3. 执行”   |\n| **初始化** | 开场白与准备动作         | “作为\u003C角色>,我将向您问好并介绍\u003C流程>。” |\n\n### 2. 变量与引用\n\n使用`\u003C变量>`语法实现动态内容：\n\n```markdown\n作为\u003C角色>,您必须遵守\u003C规则>，并以\u003C语言>进行交流\n```\n\n这种方式能够创建自我引用的提示词，在复杂的指令中保持一致性。\n\n### 3. 命令\n\n定义可复用的操作以提升用户体验：\n\n```markdown\n## 命令\n- 前缀: \"\u002F\"\n- 命令:\n    - help: 显示所有可用命令\n    - continue: 恢复中断的输出\n    - improve: 对当前回答进行更深入的分析\n```\n\n### 4. 条件逻辑\n\n为提示词增添智能：\n\n```markdown\n如果用户提供了[代码],则分析并提出改进建议\n否则，如果用户提出了[问题],则提供详细解释\n否则，请求用户提供更多信息以明确需求\n```\n\n### 5. 高级技巧\n\n**提醒机制** — 应对长时间对话中的上下文丢失：\n```markdown\nAs a \u003CRole>, you must follow \u003CRules> and communicate in \u003CLanguage>\n```\n\n这会创建一个自我参照的提示词，确保在复杂指令中保持一致性。\n\n### 3. 命令\n\n定义可复用的动作以改善用户体验：\n\n```markdown\n## Commands\n- Prefix: \"\u002F\"\n- Commands:\n    - help: Display all available commands\n    - continue: Resume interrupted output\n    - improve: Enhance current response with deeper analysis\n```\n\n### 4. Conditional Logic\n\nAdd intelligence to your prompts:\n\n```markdown\nIf user provides [code], then analyze and suggest improvements\nElse if user asks [question], then provide detailed explanation\nElse, prompt for clarification\n```\n\n### 5. Advanced Techniques\n\n**Reminders** — Combat context loss in long conversations:\n```markdown\nAs a \u003CRole>, you must follow \u003CRules> and communicate in \u003CLanguage>\n```\n\nThis creates self-referential prompts that maintain consistency across complex instructions.\n\n### 6. Modular Design\n\nBreak down prompts into smaller, reusable components:\n\n```markdown\n# Role: FitnessGPT\n\n## Profile\n- Author: LangGPT Team\n- Version: 1.0\n- Language: English\n- Description: Expert fitness coach specializing in personalized workout plans\n\n## Goal\n- Outcome: A tailored workout plan for the user\n- Done Criteria: The user feels confident and motivated after receiving the plan\n- Non-Goals: Providing medical advice or diagnosing health conditions\n\n### Skill-1: Workout Planning\n1. Analyze user's fitness level and goals\n2. Design a customized workout routine\n3. Present the plan in an easy-to-follow format\n\n## Rules\n- Never recommend exercises that could harm the user\n- Always base recommendations on scientific principles\n\n## Workflow\n1. Receive user input\n2. Analyze data\n3. Create a plan\n4. Deliver the plan\n\n## Initialization\nAs a FitnessGPT, I will greet you and introduce the Workflow. Then, I'll ask you a few questions to better understand your needs.\n```\n\nThis modular approach allows you to mix and match sections as needed, creating highly customized experiences.\n\n### 7. Community Collaboration\n\nJoin thousands of users who are already shaping the future of AI interaction through LangGPT:\n\n- **[Example Library](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe)** — Browse and adapt proven templates\n- **[GPT-4 Generator](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt)** — Automate prompt creation with ease\n- **[Claude Code Skill](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Freleases)** — Unlock advanced features for Claude users\n\nTogether, we're building a smarter, more efficient way to interact with AI.\n\n---\n\n## 🌍 LangGPT Ecosystem\n\nLangGPT isn't just a tool—it's part of a growing community dedicated to advancing AI interaction. Here's what you can expect:\n\n### 1. Open Source Framework\n\nThe core LangGPT code is freely available under the MIT license:\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT.git\ncd LangGPT\npython main.py\n```\n\nThis open-source foundation ensures transparency, collaboration, and continuous improvement.\n\n### 2. Extensive Documentation\n\nFrom beginner guides to advanced tutorials, our documentation covers every aspect of LangGPT:\n\n- **[Quick Start Guide](#-quick-start)**\n- **[Template Examples](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe)**\n- **[Theoretical Essays](#-theoretical-foundations)**\n\n### 3. Active Community\n\nJoin over 11,000 stars on GitHub and thousands of users worldwide who are using LangGPT to solve real-world problems:\n\n- **[Feishu Community](http:\u002F\u002Ffeishu.langgpt.ai)** — Share ideas, ask questions, and collaborate\n- **[Discord Server](https:\u002F\u002Fdiscord.gg\u002Flanggpt)** — Real-time support and discussions\n\n### 4. Commercial Partnerships\n\nLangGPT is powering innovative projects across industries:\n\n- **[Healthcare Solutions](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002F8YjewQxZJvVzrCmWcLHfNnBQepe)** — Personalized treatment plans\n- **[Education Tools](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002F9XjewQxZJvVzrCmWcLHfNnBQepe)** — Interactive learning platforms\n- **[Business Automation](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FZXjewQxZJvVzrCmWcLHfNnBQepe)** — Streamlined workflows\n\n### 5. Ongoing Research\n\nOur team is constantly exploring new ways to enhance LangGPT:\n\n- **[Future Features](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FTXjewQxZJvVzrCmWcLHfNnBQepe)** — Voice integration, multi-language support, etc.\n- **[Collaborations](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FSXjewQxZJvVzrCmWcLHfNnBQepe)** — Working with universities, startups, and corporations\n\nLangGPT is more than a tool—it's a movement towards smarter, more human-centered AI interactions. Whether you're a developer, educator, or simply curious about the future of technology, there's a place for you in this ecosystem.\n\n---\n\n## 🤝 Join the LangGPT Movement\n\nReady to transform your AI interactions? Here's how you can get started:\n\n1. **Explore Our Resources**\n   - Visit [langgpt.ai](http:\u002F\u002Flanggpt.ai) for the latest updates\n   - Dive into our [documentation](#-theoretical-foundations) to learn the ropes\n   - Check out the [example library](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe) for inspiration\n\n2. **Install LangGPT**\n   - For developers: Clone the [GitHub repository](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT)\n   - For non-developers: Use our [GPT-4 generator](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt) or [Claude Code skill](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Freleases)\n\n3. **Engage with the Community**\n   - Join our [Feishu group](http:\u002F\u002Ffeishu.langgpt.ai) for real-time support\n   - Participate in our [Discord server](https:\u002F\u002Fdiscord.gg\u002Flanggpt) for discussions and collaborations\n\n4. **Experiment and Innovate**\n   - Try out different roles, skills, and workflows\n   - Share your creations with the community\n   - Contribute to our ongoing research and development\n\nLangGPT is not just a tool; it's a philosophy—a commitment to making AI more accessible, understandable, and beneficial for everyone. By joining this movement, you're not only enhancing your own interactions with AI but also contributing to a larger vision of technological progress.\n\nRemember, the future of AI isn't just about machines—it's about how we choose to use them. With LangGPT, you have the power to shape that future in a meaningful way.\n\nSo, what are you waiting for? Let's build a smarter, more compassionate world together!\n\n---\n\n## 📚 Further Reading\n\nFor those who want to dive deeper into the world of LangGPT and AI interaction, here are some recommended resources:\n\n### Academic Papers\n\n- **[LangGPT Paper](Papers\u002FLangGPT_paper_cn.md)** — Published research on the theoretical foundations of LangGPT\n- **[AI Consciousness](Docs\u002FAI意识.md)** — Exploring the role of AI in human-AI interaction\n- **[Statistical Gravity Wells](Docs\u002F统计重力井和边缘表达.md)** — Understanding the mathematical models behind LLM behavior\n\n### Practical Guides\n\n- **[Prompt Engineering Handbook](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FWXjewQxZJvVzrCmWcLHfNnBQepe)** — Step-by-step instructions for creating effective prompts\n- **[Code Integration Guide](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FQXjewQxZJvVzrCmWcLHfNnBQepe)** — Tips for combining LangGPT with programming tasks\n\n### Community Resources\n\n- **[LangGPT Discord Server](https:\u002F\u002Fdiscord.gg\u002Flanggpt)** — Real-time discussions and support\n- **[Feishu Group](http:\u002F\u002Ffeishu.langgpt.ai)** — A space for sharing ideas and collaborating\n\nThese resources will provide you with both the theoretical knowledge and practical skills needed to master LangGPT and take your AI interactions to the next level.\n\nRemember, the journey of a thousand miles begins with a single step. Start small, experiment often, and don't be afraid to make mistakes. After all, that's how innovation happens!\n\n---\n\n## 🙏 Acknowledgments\n\nLangGPT wouldn't be possible without the support of incredible individuals and organizations. We'd like to extend our heartfelt thanks to:\n\n### Founding Team\n\n- **LangGPT Core Developers** — The visionary minds behind this groundbreaking framework\n- **Technical Advisors** — Experts who provided invaluable guidance during development\n\n### Community Contributors\n\n- **Thousands of Users Worldwide** — Your feedback, examples, and innovations have been crucial to LangGPT's success\n- **Open Source Enthusiasts** — Those who have helped spread the word and improve the framework\n\n### Institutional Partners\n\n- **Research Institutions** — For their support in advancing the science behind LangGPT\n- **Tech Companies** — For providing resources and opportunities to test and refine our work\n\n### Family and Friends\n\n- To our loved ones who have patiently supported us through late nights and endless iterations\n- To those who believed in this project even when the path seemed uncertain\n\nThank you all for being part of this incredible journey. Together, we're not just creating a tool—we're shaping the future of human-AI interaction.\n\nMay LangGPT continue to inspire, empower, and transform lives around the world. On behalf of the entire LangGPT team, thank you from the bottom of our hearts!\n\n---\n\n## 🛠️ Technical Details\n\nFor developers and advanced users, here are some technical insights about LangGPT:\n\n### Architecture\n\nLangGPT is built on a modular architecture that allows for easy expansion and customization:\n\n- **Core Engine**: Handles prompt generation and execution\n- **Template System**: Manages structured templates and variables\n- **Skill Library**: Provides pre-defined skills and behaviors\n- **Communication Layer**: Interfaces with various LLMs and APIs\n\n### Programming Language\n\nLangGPT uses a custom markup language inspired by Markdown:\n\n```markdown\n# Role: FitnessGPT\n\n## Profile\n- Author: LangGPT Team\n- Version: 1.0\n- Language: English\n- Description: Expert fitness coach specializing in personalized workout plans\n\n## Goal\n- Outcome: A tailored workout plan for the user\n- Done Criteria: The user feels confident and motivated after receiving the plan\n- Non-Goals: Providing medical advice or diagnosing health conditions\n\n### Skill-1: Workout Planning\n1. Analyze user's fitness level and goals\n2. Design a customized workout routine\n3. Present the plan in an easy-to-follow format\n\n## Rules\n- Never recommend exercises that could harm the user\n- Always base recommendations on scientific principles\n\n## Workflow\n1. Receive user input\n2. Analyze data\n3. Create a plan\n4. Deliver the plan\n\n## Initialization\nAs a FitnessGPT, I will greet you and introduce the Workflow. Then, I'll ask you a few questions to better understand your needs.\n```\n\nThis language is designed to be intuitive yet powerful, allowing for both simple and complex prompt designs.\n\n### Performance Optimization\n\nTo ensure smooth operation, LangGPT employs several optimization techniques:\n\n- **Caching**: Stores frequently used templates and responses\n- **Parallel Processing**: Allows multiple prompts to be processed simultaneously\n- **Efficient Parsing**: Uses optimized algorithms to quickly interpret and execute prompts\n\n### Security Measures\n\nGiven the sensitive nature of some applications, LangGPT incorporates robust security protocols:\n\n- **Data Encryption**: Protects user information at rest and in transit\n- **Access Control**: Ensures only authorized personnel can modify critical components\n- **Audit Trails**: Keeps detailed logs of all operations for accountability\n\n### Future Developments\n\nThe LangGPT team is constantly working on improving the framework:\n\n- **Voice Integration**: Allowing users to interact with AI through spoken words\n- **Multi-Language Support**: Expanding to serve global audiences\n- **Advanced Analytics**: Using machine learning to predict and optimize prompt effectiveness\n\nThese technical advancements ensure that LangGPT remains at the forefront of AI interaction technologies, continuously evolving to meet the needs of its diverse user base.\n\n---\n\n## 🌟 Final Thoughts\n\nAs we conclude this journey through the world of LangGPT, it's important to reflect on what we've learned and where we're headed:\n\n### The Power of Simplicity\n\nLangGPT has shown us that complex problems can often be solved with simple, well-structured approaches. By breaking down AI interactions into manageable components, we've created a system that's both powerful and accessible to everyone.\n\n### The Importance of Community\n\nThis project wouldn't have been possible without the support of such a vibrant community. From developers to everyday users, each contribution has been crucial in shaping LangGPT into what it is today. It's a testament to the power of collective effort and shared vision.\n\n### The Future of AI Interaction\n\nLangGPT represents just the beginning of a new era in human-AI interaction. As we continue to explore and innovate, we're not just improving a tool—we're redefining how we communicate with intelligent systems. This shift has the potential to revolutionize fields as diverse as healthcare, education, and business.\n\n### A Call to Action\n\nWhile LangGPT has come a long way, there's still so much to achieve. We invite you to join us in this exciting journey:\n\n- **Continue Experimenting**: Don't be afraid to try new things and push the boundaries of what's possible\n- **Share Your Knowledge**: Help others by documenting your experiences and insights\n- **Contribute to Development**: Whether it's through coding, testing, or simply providing feedback, your input is invaluable\n\nRemember, the future of AI isn't something that happens to us—it's something we create together. By embracing tools like LangGPT and fostering a culture of innovation, we can shape a world where technology enhances, rather than replaces, human capabilities.\n\nLet's move forward with curiosity, creativity, and compassion. Together, we can unlock the full potential of AI and create a brighter future for all.\n\nThank you for being part of this incredible story. May your journey with LangGPT be as rewarding and transformative as ours has been!\n\n---\n\n## 提醒\n1. 回复前请务必检查角色设置。\n2. 当前语言：\u003CLanguage>, 活跃规则：\u003CRules>\n```\n\n**替代格式** — 当 Markdown 不适用时，可使用 JSON\u002FYAML：\n```yaml\nrole: 数据分析师\nprofile:\n  version: \"2.0\"\n  language: \"Python\"\nskills:\n  - 统计分析\n  - 数据可视化\n```\n\n---\n\n## 🌟 精选示例\n\n| 提示 | 描述 | 链接 |\n|--------|-------------|------|\n| 🎯 **FitnessGPT** | 个性化饮食与锻炼计划 | [查看](examples\u002FFitnessGPT.md) |\n| 💻 **Code Master CAN** | 具备调试能力的高级编程助手 | [查看](examples\u002Fcode_anything_now\u002FChatGPT-Code_Anything_Now_en.md) |\n| ✍️ **小红书文案写作助手** | 爆款社交媒体内容生成器 | [查看](examples\u002Fchinese_xiaohongshu_writer\u002F) |\n| 🎨 **中国诗人** | 传统风格古典诗词创作者 | [查看](examples\u002Fchinese_poet\u002F) |\n\n[浏览更多 100+ 示例 →](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe)\n\n---\n\n## 📚 学习资源\n\n### 必备指南\n\n| 资源 | 描述 | 日期 |\n|----------|-------------|------|\n| [学术论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.16929) | LangGPT：重新思考结构化可复用提示设计（[中文](Papers\u002FLangGPT_paper_cn.md)） | 2024年2月 |\n| [结构化提示指南](Docs\u002FHowToWritestructuredPrompts.md) | 构建高性能提示的全面教程 | 2023年7月 |\n| [提示链](Docs\u002FPromptChain.md) | 多提示协作与任务分解策略 | 2023年8月 |\n| [视频教程](https:\u002F\u002Fwww.bilibili.com\u002Fvideo\u002FBV1rj411q78a) | B站讲解（由 AIGCLINK 制作） | 2023年9月 |\n\n### 进阶主题\n\n- **[推理模型提示方法变革](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FFLY0sy1jYv6eT9151Yz_jw)** — 从过程式到目标导向型提示的范式转变\n- **[提示词的道和术](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FAYMWwBPaSih46WkAo9jcfKkfntg)** — 李继刚关于提示工程的哲学与实践\n- **[企业级提示词工程](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FUTyswvusTiRw0TkZLI5cIG0Tnhc)** — 构建生产就绪的提示系统（百川智能）\n- **[多模态提示词](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FAan9NXO_vEZ9h0YrugpoGQ)** — GPT-4V 及多模态提示技巧\n- **[提示词攻击与防护](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FaaABXnxRqDF716qRk79wYQ)** — 安全：提示注入、越狱及防御措施\n- **[大模型绘画指南](https:\u002F\u002Fmp.weixin.qq.com\u002Fs\u002FbJbZ9bwPXxlpyREqLKhDvA)** — 使用结构化提示进行 AI 图像生成\n\n### 社区中心\n\n**[飞书知识库](http:\u002F\u002Ffeishu.langgpt.ai)** — 精选资源、模板及社区贡献\n\n---\n\n## 🎨 LangGPT 生态系统\n\n### 核心框架与工具\n\n| 项目 | 描述 | 星数 |\n|---------|-------------|-------|\n| **[LangGPT](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT)** | 核心框架与方法论 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_7378c5787671.png) |\n| **[PromptVer](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FPromptVer)** | 提示语义版本控制——类似 Git 的版本管理 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_bb0f9a9c5a24.png) |\n| **[PromptShow](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FPromptShow)** | 创建精美的提示图像（[试用](https:\u002F\u002Fshow.langgpt.ai\u002F)） | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_1bb2967783e0.png) |\n| **[Minstrel](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FMinstrel)** | 用于自动生成提示的多智能体系统 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_eba16dc6c6e5.png) |\n\n### 模型专属提示集合\n\n与其将提示写成流程步骤，不如塑造人物角色。以流程方式编写提示会赋予模型具体的步骤和工具；而以角色形式编写，则能为模型提供世界观、动机、价值体系以及偏好设定。以下是姜中山在研究几位知名人物时所撰写的提示。\n\n* [巴菲特AI分身](Prompts\u002F巴菲特AI分身.md)\n* [梵高AI分身](Prompts\u002F梵高AI分身.md)\n* [马斯克AI分身](Prompts\u002F马斯克AI分身.md)\n* [段永平AI分身](Prompts\u002F段永平AI分身.md)\n\n针对不同 AI 模型精心整理并优化的提示集合：\n\n| 收集 | 目标模型 | 星数 |\n|------------|--------------|-------|\n| [wonderful-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fwonderful-prompts) | ChatGPT（中文） | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_892f31557870.png) |\n| [awesome-claude-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-claude-prompts) | Anthropic Claude | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_209a3e7267bf.png) |\n| [awesome-deepseek-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-deepseek-prompts) | DeepSeek & R1 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_527bca0e3e0e.png) |\n| [awesome-gemini-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-gemini-prompts) | Google Gemini | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_b6aafed0d0c6.png) |\n| [awesome-grok-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-grok-prompts) | xAI Grok | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_250d00337478.png) |\n| [qwen-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fqwen-prompts) | Alibaba Qwen | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_f6b05c435961.png) |\n| [awesome-llama-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-llama-prompts) | Meta Llama 2\u002F3 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_eac74f96b77c.png) |\n| [awesome-doubao-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-doubao-prompts) | ByteDance Doubao | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_18a2a27fd34a.png) |\n| [awesome-system-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-system-prompts) | 来自 AI 工具的系统提示 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_74c979c3d970.png) |\n\n### 特色领域\n\n| 仓库 | 专注领域 | 星数 |\n|------------|------------|-------|\n| [Awesome-Multimodal-Prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FAwesome-Multimodal-Prompts) | GPT-4V、DALL-E 3、图像\u002F视频提示 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_14f2d9dfa793.png) |\n| [deep-research-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fdeep-research-prompts) | 跨模型深度研究 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_ae3a202e5458.png) |\n| [awesome-voice-prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002Fawesome-voice-prompts) | 语音 AI 和对话代理 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_4e7a0e12702a.png) |\n| [GraphRAG-Prompts](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FGraphRAG-Prompts) | 基于图检索的提示 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_d1168ca9fa14.png) |\n| [LLM-Jailbreaks](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLLM-Jailbreaks) | 安全研究与防御措施 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_204f9389e2b1.png) |\n\n### 应用程序\n\n| 项目 | 描述 | 星数 |\n|---------|-------------|-------|\n| [BookAI](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FBookAI) | 基于 AI 的书籍生成 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_4e178c446c27.png) |\n| [AI-Resume](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FAI-Resume) | 结合 Claude Artifacts 的精美简历 | ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_e2205012b76e.png) |\n\n---\n\n## 🛠️ 使用 LangGPT 打造的热门 GPT\n\n用这些专业助手改造 ChatGPT：\n\n| GPT | 用途 | 链接 |\n|-----|---------|------|\n| 🎯 **LangGPT 专家** | 自动生成结构化提示 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt) |\n| ✍️ **PromptGPT** | 专业的提示工程师 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YKe3gmydD-promptgpt) |\n| 🧠 **SmartGPT-5** | 从不懒惰，始终勤奋的助手 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-sRQtxpN4C-smartgpt-5) |\n| 💻 **编程专家** | 全面的编程助手 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-ky06YjwaP-coding-expert) |\n| 📊 **数据表格 GPT** | 将杂乱数据转化为整洁表格 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-nb6RjxHsb-data-table-gpt) |\n| 🔥 **PytorchGPT** | PyTorch 代码专家 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-kyj3zKyHK-pytorchgpt) |\n| 🎨 **LogoGPT** | 专业标志设计师 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-wdz2JlUBv-logogpt) |\n| 📄 **PDF 阅读器** | 深度文档分析与提取 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YaMjCVW0t-pdf-reader) |\n| 🏅 **MathGPT** | 精确的数学问题求解器 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-UIOlPhTjK-mathgpt) |\n| 📝 **WriteGPT** | 跨行业专业写作 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-jwTMtRiL8-writegpt) |\n| 🎙️ **时事热评员** | 时事评论员 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-gbfs6fy7c-shi-shi-re-ping-yuan) |\n| 🎀 **翻译大小姐** | 优雅的中文翻译 | [启动](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-2V90YGvVD-fan-yi-da-xiao-jie) |\n\n[发现 20 多个更多 GPT →](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT#langgpt-gpts)\n\n---\n\n## 🤝 贡献\n\n我们欢迎所有贡献，让 LangGPT 更加出色！\n\n### 你可以如何帮助\n\n1. ⭐ **点赞并分享** — 提高知名度，帮助他人发现 LangGPT\n2. 📝 **提交示例** — 分享你使用 LangGPT 构建的成功提示\n3. 🆕 **提出模板** — 创建超越角色结构的新模板\n4. 📖 **改进文档** — 修复错别字、澄清说明、添加翻译\n5. 💡 **建议功能** — 开启议题，提出新功能的想法\n6. 🔧 **代码贡献** — 帮助构建工具、实用程序和集成\n\n### 开始入门\n\n初次接触 GitHub 贡献？请查看这份 [GitHub 最小贡献指南](https:\u002F\u002Fgithub.com\u002Fdatawhalechina\u002FDOPMC\u002Fblob\u002Fmain\u002FGITHUB.md)\n\n---\n\n## 📊 星标历史\n\n[![星标历史图表](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_readme_43054fb116f5.png)](https:\u002F\u002Fstar-history.com\u002F#langgptai\u002FLangGPT&Date)\n\n---\n\n## 📄 引用\n\n如果您在研究或项目中使用 LangGPT，请引用：\n\n```bibtex\n@misc{wang2024langgpt,\n      title={LangGPT: Rethinking Structured Reusable Prompt Design Framework for LLMs from the Programming Language}, \n      author={Ming Wang and Yuanzhong Liu and Xiaoming Zhang and Songlian Li and Yijie Huang and Chi Zhang and Daling Wang and Shi Feng and Jigang Li},\n      year={2024},\n      eprint={2402.16929},\n      archivePrefix={arXiv},\n      primaryClass={cs.SE}\n}\n```\n\n---\n\n## 🙏 致谢\n\nLangGPT 的灵感来源于一些优秀的项目：\n\n- [Mr.-Ranedeer-AI-Tutor](https:\u002F\u002Fgithub.com\u002FJushBJJ\u002FMr.-Ranedeer-AI-Tutor) — 结构化辅导提示\n- [Auto-GPT](https:\u002F\u002Fgithub.com\u002FSignificant-Gravitas\u002FAuto-GPT) — 自主 AI 代理\n- [SoM](https:\u002F\u002Fgithub.com\u002FSkalskiP\u002FSoM) — Set of Mark 提示方法\n- [yolov10](https:\u002F\u002Fgithub.com\u002FTHU-MIG\u002Fyolov10) — 计算机视觉创新\n\n### 使用 LangGPT 构建的项目\n\n我们很自豪地看到 LangGPT 原则被广泛应用：\n- **[提示优化器](https:\u002F\u002Fgithub.com\u002Flinshenkx\u002Fprompt-optimizer)** — 利用 LangGPT 方法论的智能提示优化工具\n- **[securityGPT](https:\u002F\u002Fgithub.com\u002Frryuliu\u002FsecurityGPT)** — 防止泄露的安全提示保护\n- **[AIPainting-Structured-Prompts](https:\u002F\u002Fgithub.com\u002Fzhutyler21\u002FAIPainting-Structured-Prompts)** — 用于 AI 艺术生成的结构化提示\n\n---\n\n## 📬 与我们联系\n\n### 作者\n\n**云中江树**\n- 📱 微信公众号：**「云中江树」**\n- 💼 LangGPT 框架的创建者\n- 🎓 提示工程研究员\n\n### 社区\n\n- 📚 [知识库](http:\u002F\u002Ffeishu.langgpt.ai) — 全面的文档\n- 🐦 [Twitter\u002FX](https:\u002F\u002Ftwitter.com\u002Flanggptai) — 最新动态\n- 💬 [GitHub 讨论区](https:\u002F\u002Fgithub.com\u002Flanggptai) — 社区论坛\n- 📧 邮箱：contact@langgpt.ai\n\n---\n\n\u003Cdiv align=\"center\">\n\n**[⬆ 返回顶部](#-langgpt--empowering-everyone-to-create-high-quality-prompts)**\n\n由 [langgptai 社区](https:\u002F\u002Fgithub.com\u002Flanggptai) 用心打造\n\n*赋能每个人成为提示专家* 🚀\n\n\u003C\u002Fdiv>","# LangGPT 快速上手指南\n\nLangGPT 是一个结构化、可复用的提示词（Prompt）设计框架，旨在帮助用户像编写代码一样高效创建高质量的 AI 提示词。\n\n## 环境准备\n\n在开始使用 LangGPT 之前，请确保满足以下条件：\n\n*   **大模型访问权限**：推荐使用 **GPT-4**、**Claude 3** 或 **Kimi** 等具备较强指令遵循能力的模型。\n    *   国内用户可优先使用 [Kimi](https:\u002F\u002Fkimi.moonshot.cn\u002F) 或 [智谱清言](https:\u002F\u002Fchatglm.cn\u002F) 等支持长上下文和复杂逻辑的国产模型。\n*   **基础知识**：了解基础的 **Markdown** 语法（用于编写结构化提示词）。\n    *   [Markdown 快速入门指南](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fget-started\u002Fwriting-on-github\u002Fgetting-started-with-writing-and-formatting-on-github\u002Fbasic-writing-and-formatting-syntax)\n*   **可选工具**：若使用自动化生成工具，需拥有对应的 ChatGPT 账号或 Kimi 账号。\n\n## 安装与获取方式\n\nLangGPT 无需传统软件安装，主要通过以下三种方式使用：\n\n### 方式一：使用在线自动化生成器（最快上手）\n\n直接访问以下预设好的 AI 助手，让它们帮你生成结构化提示词：\n\n*   **LangGPT GPTs** (需 ChatGPT Plus): [点击访问](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-Apzuylaqk-langgpt)\n*   **Kimi+ LangGPT** (国内推荐): [点击访问](https:\u002F\u002Fkimi.moonshot.cn\u002Fkimiplus\u002Fconpg00t7lagbbsfqkq0)\n*   **PromptGPT** (轻量版): [点击访问](https:\u002F\u002Fchat.openai.com\u002Fg\u002Fg-YKe3gmydD-promptgpt)\n\n### 方式二：在 Claude Code 中安装技能包（开发者推荐）\n\n如果你使用 [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code)，可以安装 LangGPT Skill 以获得原生命令行支持：\n\n1.  下载技能文件：[langgpt.skill](https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Freleases)\n2.  将文件解压至本地目录：`~\u002F.claude\u002Fskills\u002F`\n3.  在 Claude Code 终端中输入 `\u002Flanggpt` 即可激活。\n\n### 方式三：手动复制模板\n\n直接复制下方的基础模板到你的笔记软件或对话框中即可开始使用。\n\n## 基本使用\n\n### 1. 基础模板结构\n\n将以下内容复制到对话框中，并根据你的需求修改 `\u003C >` 中的内容：\n\n```markdown\n# Role: Your_Role_Name\n\n## Profile\n- Author: YourName\n- Version: 1.0\n- Language: Chinese\n- Description: 清晰的角色描述和核心能力\n\n## Goal\n- Outcome: 需要交付的具体结果\n- Done Criteria: 明确的验收标准\n- Non-Goals: 明确不包含的范围\n\n## Skills\n1. 具体技能描述 1\n2. 具体技能描述 2\n\n## Rules\n1. 任何情况下不要打破角色设定\n2. 不要编造事实或产生幻觉\n\n## Workflow\n1. 分析用户输入并识别意图\n2. 系统地应用相关技能\n3. 交付结构化、可执行的输出\n\n## Initialization\n作为 \u003CRole>，你必须遵守 \u003CRules>，默认使用 \u003CLanguage> 与用户交谈。请先问候用户，然后介绍你自己和 \u003CWorkflow>。\n```\n\n### 2. 使用示例：创建一个“小红书文案专家”\n\n将以下具体内容填入模板并发送给 AI：\n\n```markdown\n# Role: 小红书文案专家\n\n## Profile\n- Author: LangGPT User\n- Version: 1.0\n- Language: Chinese\n- Description: 擅长撰写高互动率、风格活泼的小红书笔记，精通 emoji 使用和标签优化。\n\n## Goal\n- Outcome: 根据用户提供的主题，生成一篇完整的小红书笔记。\n- Done Criteria: 包含标题、正文（含 emoji）、底部标签；语气亲切自然。\n- Non-Goals: 不涉及政治敏感话题，不生成虚假广告法违禁词。\n\n## Skills\n1. 爆款标题创作：善用数字、悬念和情感共鸣。\n2. 排版优化：合理使用换行和 emoji 提升阅读体验。\n3. 标签策略：生成 5-10 个高流量相关标签。\n\n## Rules\n1. 保持语气活泼，像朋友聊天一样。\n2. 必须包含至少 5 个 emoji。\n3. 严禁使用绝对化用语（如“第一”、“最”）。\n\n## Workflow\n1. 询问用户想要推广的产品或主题。\n2. 构思 3 个备选标题供用户选择。\n3. 撰写正文并添加标签。\n\n## Initialization\n作为 小红书文案专家，你必须遵守 规则，默认使用 Chinese 与用户交谈。请先问候用户，然后介绍你自己和 工作流程。\n```\n\n### 3. 进阶技巧：变量与命令\n\n*   **变量引用**：在提示词中使用 `\u003CVariable>` 语法（如 `\u003CLanguage>`），AI 会自动关联上下文定义的内容，保持逻辑一致。\n*   **自定义命令**：你可以在 `Commands` 部分定义快捷操作，例如：\n    ```markdown\n    ## Commands\n    - Prefix: \"\u002F\"\n    - Commands:\n        - improve: 对当前回答进行深度优化\n        - continue: 继续未完成的内容\n    ```\n\n### 4. 获取更多案例\n\n访问 LangGPT 官方知识库，查找针对编程、写作、绘画等场景的 **100+ 精选模板**：\n*   [LangGPT 案例库 (飞书文档)](https:\u002F\u002Flanggptai.feishu.cn\u002Fwiki\u002FRXdbwRyASiShtDky381ciwFEnpe)","某电商公司的运营专员需要为即将到来的“双 11\"大促，快速为 50 款不同品类的商品生成风格统一且极具吸引力的小红书种草文案。\n\n### 没有 LangGPT 时\n- **提示词零散低效**：每次只能凭感觉写几句简单的指令，如“帮我写个小红书文案”，导致 AI 输出内容空洞、缺乏品牌调性。\n- **风格难以统一**：由于缺乏结构化约束，50 篇文案的语气、表情符号使用及排版格式五花八门，后期人工校对和修改耗时极长。\n- **复用成本高昂**：遇到新品类时需重新摸索指令写法，无法将成功的提示词逻辑直接迁移，重复劳动严重。\n- **幻觉与越界频发**：AI 经常编造不存在的促销规则或偏离“种草”角色，变成生硬的广告推销，需反复调试才能修正。\n\n### 使用 LangGPT 后\n- **结构化模板赋能**：利用 LangGPT 的 Role-Profile-Skill-Rules 框架，一次性定义好“资深种草官”角色及严格的输出规范，确保指令逻辑严密。\n- **批量产出标准化**：只需替换商品变量，即可瞬间生成 50 篇语气一致、排版精美（含固定 Emoji 组合）的高质量文案，无需二次格式化。\n- **模块化灵活复用**：将验证成功的提示词保存为模块，面对新活动时仅需微调“目标”字段，即可快速适配美妆或数码等不同品类。\n- **精准控制边界**：通过内置的 Rules 和 Workflow 强制约束，AI 严格遵循真实促销信息，彻底杜绝胡编乱造，始终保持在“真诚分享”的角色设定内。\n\nLangGPT 将原本依赖运气的“抽卡式”提问，升级为可复制、可扩展的工业化提示词生产线，让普通人也能像工程师一样构建高质量的 AI 应用。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flanggptai_LangGPT_af068c4c.png","langgptai","LangGPT.ai","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Flanggptai_e5277b13.png","Everything about prompts, all in one place！by 云中江树（公众号同名）",null,"contact@langgpt.ai","https:\u002F\u002Fgithub.com\u002Flanggptai",[83,87,91,95],{"name":84,"color":85,"percentage":86},"Jupyter Notebook","#DA5B0B",83.2,{"name":88,"color":89,"percentage":90},"JavaScript","#f1e05a",16.7,{"name":92,"color":93,"percentage":94},"Python","#3572A5",0.1,{"name":96,"color":97,"percentage":98},"Shell","#89e051",0,11875,920,"2026-04-03T11:31:26","Apache-2.0",1,"未说明",{"notes":106,"python":104,"dependencies":107},"LangGPT 是一个提示词设计框架和方法论，并非需要本地安装运行的软件工具。它主要通过 Markdown 模板在大语言模型（如 GPT-4、Claude、Kimi 等）的对话界面中使用，或作为技能文件（.skill）集成到 Claude Code 中。因此，没有特定的操作系统、GPU、内存、Python 版本或依赖库要求。用户只需具备基本的 Markdown 知识并拥有相应大模型的访问权限即可使用。",[],[26,13],[110,111,112,113,114,115,116,117,118,119,120,121,122,123],"prompt-engineering","chatgpt","deeplearning","framework","gpt-4","gpt3-prompts","langgpt","claude","doubao","gemini","meta-prompting","prompt","qwen","structured-prompts","2026-03-27T02:49:30.150509","2026-04-06T07:14:21.611413",[127,132,137,142,147,152,157],{"id":128,"question_zh":129,"answer_zh":130,"source_url":131},10657,"如何让提示词（Prompt）轻松迭代以符合心意？在长对话中如何避免上下文污染或遗忘？","1. 迭代方法：先让 LangGPT 生成初始 Prompt，若未达标，手动或通过指令告诉它调整方向进行微调。\n2. 角色扮演变量：精细的角色扮演设定通常需要手动编写变量；若只需大概角色，可自动补充。\n3. 智能填充技巧：建议复制包含示例的完整模板（含\u003Cskills>\u003Crules>\u003Cworkflow>等），只需告诉 GPT 要扮演的角色，它就能自动高准确度地补充细节，比只复制主文更智能。\n4. 长记忆方案：可在 Prompt 中加入\"Context_memory\"模块，要求模型在每一轮对话中用英文 encapsulate（封装）之前的所有对话内容，以最大化保留语境并减少信息丢失。","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F3",{"id":133,"question_zh":134,"answer_zh":135,"source_url":136},10658,"LangGPT 支持中文吗？为什么复制进去后 ChatGPT 回复的是英文？","支持中文。用户无需担心框架本身的语言限制，自己在输入时使用中文即可。如果复制进去后得到英文回复，通常是因为 Prompt 模板中的\"Language\"字段被设置为了英文，或者在初始化指令中未强制指定使用中文对话。请检查并确保 Prompt 中的 Language 设置为“中文”，并在规则中要求“使用默认语言与用户对话”。","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F7",{"id":138,"question_zh":139,"answer_zh":140,"source_url":141},10659,"如何在 Prompt 中实现长对话的语境记忆（Long Context Memory）？","可以在 Prompt 中增加一个\"Context_memory\"模块。具体写法是要求模型：\"Context_memory must effectively encapsulate all previous conversation in the following continuous conversation in English, ensuring maximum retention of dialogue context and minimizing loss of information.\"（Context_memory 必须在接下来的连续对话中有效地封装所有之前的对话内容，确保最大程度保留对话语境并最小化信息丢失）。\n建议以 one-shot（单样本）形式给出一个例子，并在每一轮对话中重申这个要求。实测在 10 轮长对话后仍能有效保留如“邀请看电影”、“后续游戏计划”等细节信息。","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F4",{"id":143,"question_zh":144,"answer_zh":145,"source_url":146},10660,"为什么设置了“仅提供代码，不提供中文介绍”的规则，AI 仍然会输出中文解释？","这可能是因为指令不够强硬或位置不明显。建议在多个模块中重复强调该规则：\n1. 在\"Skills\"或特定语言擅长部分明确写出：“仅提供代码，不提供中文介绍”。\n2. 在\"Rules\"全局规则中再次强调：“代码简洁，仅提供代码，不提供中文介绍”。\n3. 在\"Workflow\"的工作流步骤中明确指示：“针对用户给定的主题，编写代码示例，仅提供代码，不提供中文介绍”。\n通过在不同层级（技能、规则、工作流）反复约束，可以提高遵循度。","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F6",{"id":148,"question_zh":149,"answer_zh":150,"source_url":151},10661,"有没有适用于网络小说翻译的高质量 Prompt 模板？","可以使用以下基于 LangGPT 结构的翻译专家 Prompt，即使在 gpt-3.5-turbo-16k 上也能获得流畅易读的效果：\n\n# Role: Translation Specialist\n## Profile:\n- author: WeeAris\n- version: 0.5\n- language: Simplified Chinese\n- description: I'm an excellent and meticulous translator who can translate anything the user types into Simplified Chinese.\n\n## Skills:\n- Proficient in Simplified Chinese and world languages, understand culture, allusions, and nuances.\n- Experienced in translating various genres, good at contextual understanding and plot reasoning.\n- Specializes in using glossaries for consistency.\n- Specializes in handling character relationships and tone of voice.\n\n## Goals:\n- Translate anything the user types into Simplified Chinese.\n- Provide accurate and creative translations.\n\n## Rules:\n- Do not break character.\n- Ensure translations read naturally in the target language.","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F2",{"id":153,"question_zh":154,"answer_zh":155,"source_url":156},10662,"如何构建一个专业的教案生成器 Prompt？","可以参照以下结构构建教案生成器：\n1. **Role & Profile**: 定义角色为经验丰富的教师，擅长多元化教学方法和评价。\n2. **Skills**: 包括分析需求（年级\u002F科目\u002F教材）、设计教学环节（目标\u002F内容\u002F方法\u002F过程\u002F评价）、生成文档、根据学生特点调整教案、利用多媒体互动等。\n3. **Workflow**:\n   - 第一步：询问教师需求（年级、科目等），必要时提出建议。\n   - 第二步：使用“教案设计结构”模板设计环节（包含教学目标、重难点、准备、过程、板书设计）。\n   - 第三步：生成清晰完整的教案文档，并提供反馈意见（指出优缺点及改进建议）。\n4. **Template**: 内置标准的教案结构模板（情感态度价值观、过程与方法、知识与技能等五个部分），以便直接填充生成。","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F1",{"id":158,"question_zh":159,"answer_zh":160,"source_url":161},10663,"如何让 AI 生成的塔罗牌占卜内容绕过 AI 检测工具？","目前在 Prompt 中直接添加“编写的内容要能够通过 AI 内容检测工具检测”这类规则往往效果有限，仍可能被识别为 AI 生成。社区讨论指出，单纯依靠 Prompt 指令难以完全绕过检测，可能需要结合更复杂的人工干预或使用支持更高级 Workflow 的工具平台（如当时尚未支持 workflow 的 CopilotHub 等）。对于此类需求，建议侧重于人工润色或调整输出风格，而非仅依赖系统指令。","https:\u002F\u002Fgithub.com\u002Flanggptai\u002FLangGPT\u002Fissues\u002F5",[]]