[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:metagpt":3},"\u003Ch1>MetaGPT: The Multi-Agent Framework\u003C\u002Fh1>\n\u003Cp align=\"center\">\n\u003Ca rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002FFoundationAgents\u002Fmetagpt\u002FHEAD\u002Fdocs\u002Fresources\u002FMetaGPT-new-log.png\" alt=\"MetaGPT logo: Enable GPT to work in a software company, collaborating to tackle more complex tasks.\" width=\"150px\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\u003Cp align=\"center\">\n[ \u003Cb>En\u003C\u002Fb> |\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FFoundationAgents\u002Fmetagpt\u002Fblob\u002FHEAD\u002Fdocs\u002FREADME_CN.md\" rel=\"nofollow ugc noopener\">中\u003C\u002Fa> |\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FFoundationAgents\u002Fmetagpt\u002Fblob\u002FHEAD\u002Fdocs\u002FREADME_FR.md\" rel=\"nofollow ugc noopener\">Fr\u003C\u002Fa> |\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FFoundationAgents\u002Fmetagpt\u002Fblob\u002FHEAD\u002Fdocs\u002FREADME_JA.md\" rel=\"nofollow ugc noopener\">日\u003C\u002Fa> ]\n\u003Cb>Assign different roles to GPTs to form a collaborative entity for complex tasks.\u003C\u002Fb>\n\u003C\u002Fp>\u003Cp align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-blue.svg\" alt=\"License: MIT\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FDYn29wFk9z\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FJoin-Discord-gGnrXvVz7a?logo=discord\" alt=\"Discord Follow\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Ftwitter.com\u002FMetaGPT_\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Ftwitter\u002Ffollow\u002FMetaGPT?style=social\" alt=\"Twitter Follow\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\u003Ch4>\u003C\u002Fh4>\u003Ch2>News\u003C\u002Fh2>\n\u003Cp>🚀 Mar. 10, 2025: 🎉 \u003Ca href=\"https:\u002F\u002Fmgx.dev\u002F\" rel=\"nofollow ugc noopener\">mgx.dev\u003C\u002Fa> is the #1 Product of the Week on @ProductHunt! 🏆\u003C\u002Fp>\n\u003Cp>🚀 Mar.   4, 2025: 🎉 \u003Ca href=\"https:\u002F\u002Fmgx.dev\u002F\" rel=\"nofollow ugc noopener\">mgx.dev\u003C\u002Fa> is the #1 Product of the Day on @ProductHunt! 🏆\u003C\u002Fp>\n\u003Cp>🚀 Feb. 19, 2025: Today we are officially launching our natural language programming product: \u003Ca href=\"https:\u002F\u002Fmgx.dev\u002F\" rel=\"nofollow ugc noopener\">MGX (MetaGPT X)\u003C\u002Fa> - the world's first AI agent development team. More details on \u003Ca href=\"https:\u002F\u002Fx.com\u002FMetaGPT_\u002Fstatus\u002F1892199535130329356\" rel=\"nofollow ugc noopener\">Twitter\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>🚀 Feb. 17, 2025: We introduced two papers: \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fpdf\u002F2502.06855\" rel=\"nofollow ugc noopener\">SPO\u003C\u002Fa> and \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fpdf\u002F2502.12018\" rel=\"nofollow ugc noopener\">AOT\u003C\u002Fa>, check the \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FFoundationAgents\u002Fmetagpt\u002Fblob\u002FHEAD\u002Fexamples\" rel=\"nofollow ugc noopener\">code\u003C\u002Fa>!\u003C\u002Fp>\n\u003Cp>🚀 Jan. 22, 2025: Our paper \u003Ca href=\"https:\u002F\u002Fopenreview.net\u002Fforum?id=z5uVAKwmjf\" rel=\"nofollow ugc noopener\">AFlow: Automating Agentic Workflow Generation\u003C\u002Fa> accepted for \u003Cstrong>oral presentation (top 1.8%)\u003C\u002Fstrong> at ICLR 2025, \u003Cstrong>ranking #2\u003C\u002Fstrong> in the LLM-based Agent category.\u003C\u002Fp>\n\u003Cp>👉👉 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FFoundationAgents\u002Fmetagpt\u002Fblob\u002FHEAD\u002Fdocs\u002FNEWS.md\" rel=\"nofollow ugc noopener\">Earlier news\u003C\u002Fa> \u003C\u002Fp>\n\u003Ch2>Software Company as Multi-Agent System\u003C\u002Fh2>\n\u003Col>\n\u003Cli>MetaGPT takes a \u003Cstrong>one line requirement\u003C\u002Fstrong> as input and outputs \u003Cstrong>user stories \u002F competitive analysis \u002F requirements \u002F data structures \u002F APIs \u002F documents, etc.\u003C\u002Fstrong>\u003C\u002Fli>\n\u003Cli>Internally, MetaGPT includes \u003Cstrong>product managers \u002F architects \u002F project managers \u002F engineers.\u003C\u002Fstrong> It provides the entire process of a \u003Cstrong>software company along with carefully orchestrated SOPs.\u003C\u002Fstrong>\u003Col>\n\u003Cli>\u003Ccode>Code = SOP(Team)\u003C\u002Fcode> is the core philosophy. We materialize SOP and apply it to teams composed of LLMs.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002FFoundationAgents\u002Fmetagpt\u002FHEAD\u002Fdocs\u002Fresources\u002Fsoftware_company_cd.jpeg\" alt=\"A software company consists of LLM-based roles\" \u002F>\u003C\u002Fp>\n\u003Cp align=\"center\">Software Company Multi-Agent Schematic (Gradually Implementing)\u003C\u002Fp>\u003Ch2>Get Started\u003C\u002Fh2>\n\u003Ch3>Installation\u003C\u002Fh3>\n\u003Cblockquote>\n\u003Cp>Ensure that Python 3.9 or later, but less than 3.12, is installed on your system. You can check this by using: \u003Ccode>python --version\u003C\u002Fcode>.\u003Cbr \u002F>You can use conda like this: \u003Ccode>conda create -n metagpt python=3.9 && conda activate metagpt\u003C\u002Fcode>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cpre>\u003Ccode class=\"language-bash\">pip install --upgrade metagpt\n# or `pip install --upgrade git+https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT.git`\n# or `git clone https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT && cd MetaGPT && pip install --upgrade -e .`\n\u003C\u002Fcode>\u003C\u002Fpre>\n\u003Cp>\u003Cstrong>Install \u003Ca href=\"https:\u002F\u002Fnodejs.org\u002Fen\u002Fdownload\" rel=\"nofollow ugc noopener\">node\u003C\u002Fa> and \u003Ca href=\"https:\u002F\u002Fpnpm.io\u002Finstallation#using-npm\" rel=\"nofollow ugc noopener\">pnpm\u003C\u002Fa> before actual use.\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>For detailed installation guidance, please refer to \u003Ca href=\"https:\u002F\u002Fdocs.deepwisdom.ai\u002Fmain\u002Fen\u002Fguide\u002Fget_started\u002Finstallation.html#install-stable-version\" rel=\"nofollow ugc noopener\">cli_install\u003C\u002Fa>\n or \u003Ca href=\"https:\u002F\u002Fdocs.deepwisdom.ai\u002Fmain\u002Fen\u002Fguide\u002Fget_started\u002Finstallation.html#install-with-docker\" rel=\"nofollow ugc noopener\">docker_install\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch3>Configuration\u003C\u002Fh3>\n\u003Cp>You can init the config of MetaGPT by running the following command, or manually create \u003Ccode>~\u002F.metagpt\u002Fconfig2.yaml\u003C\u002Fcode> file:\u003C\u002Fp>\n\u003Cpre>\u003Ccode class=\"language-bash\"># Check https:\u002F\u002Fdocs.deepwisdom.ai\u002Fmain\u002Fen\u002Fguide\u002Fget_started\u002Fconfiguration.html for more details\nmetagpt --init-config  # it will create ~\u002F.metagpt\u002Fconfig2.yaml, just modify it to your needs\n\u003C\u002Fcode>\u003C\u002Fpre>\n\u003Cp>You can configure \u003Ccode>~\u002F.metagpt\u002Fconfig2.yaml\u003C\u002Fcode> according to the \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT\u002Fblob\u002Fmain\u002Fconfig\u002Fconfig2.example.yaml\" rel=\"nofollow ugc noopener\">example\u003C\u002Fa> and \u003Ca href=\"https:\u002F\u002Fdocs.deepwisdom.ai\u002Fmain\u002Fen\u002Fguide\u002Fget_started\u002Fconfiguration.html\" rel=\"nofollow ugc noopener\">doc\u003C\u002Fa>:\u003C\u002Fp>\n\u003Cpre>\u003Ccode class=\"language-yaml\">llm:\n  api_type: \"op\n\u003C\u002Fcode>\u003C\u002Fpre>\n",1784240407607]