MetaGPT is a multi-agent framework that assigns different roles to GPTs to form a collaborative entity for complex tasks, taking a one line requirement as input and outputting user stories, competitive analysis, requirements, data structures, APIs, and documents.
MetaGPT is a multi-agent framework that enables GPT to work in a software company, collaborating to tackle more complex tasks. It takes a one line requirement as input and outputs user stories, competitive analysis, requirements, data structures, APIs, and documents. Internally, MetaGPT includes product managers, architects, project managers, and engineers, providing the entire process of a software company along with carefully orchestrated SOPs. Its core philosophy is `Code = SOP(Team)`, materializing SOP and applying it to teams composed of LLMs.
Assigning different roles to GPTs to form a collaborative entity for complex tasks, simulating a software company as a multi-agent system that turns a one line requirement into software artifacts and code.
MetaGPT is explicitly a multi-agent framework that orchestrates LLM-based roles (product manager, architect, engineer) through carefully orchestrated SOPs to take a one-line requirement and autonomously produce code, documents, and project artifacts. The README describes an agentic workflow ("AFlow: Automating Agentic Workflow Generation") and a full software-company pipeline that runs without per-step human confirmation. Surface analysis indicates it executes arbitrary code and shell commands as engineer agents write and run generated code, consistent with multi-step autonomous action. While the entrypoint code showed no direct signals, the framework-level design and code-execution behavior place it at Tier 4.