How vivekdeshmukhrepos/multi_agent_requirement_to_code_generator works
Project type and purpose
• This is a single-service, full-stack (UI + backend logic) Streamlit application whose goal is to take plain software requirements, clarify them via a series of AI “agents,” break them into user stories, generate jQuery code for each story, then aggregate it all into one file.
• It is not a microservices architecture in the traditional sense but rather an internal multi-agent pipeline orchestrated in one Python process (app.py).
File structure
• Top-level files:
• app.py — the entry point: sets up Streamlit UI, manages the multi-agent workflow.
• requirements.txt — Python dependencies including streamlit, openai, semantic-kernel, asyncio.
• Resources/ — folder containing supplementary assets (demo video, sample inputs).
• There are no separate “frontend” or “backend” directories; the UI and orchestration live together in app.py.
• No databases, but there are external HTTP calls to the OpenAI API.
README for additional insights
• The README lists four autonomous agents: RequirementsAgent, UserStoryAgent, CodeGenAgent, AggregatorAgent.
• Technologies: Streamlit (UI), Semantic Kernel (agent framework), OpenAI GPT-4 (LLM), Async Python (concurrency).