How nahidbrur/multi-agent-competitor-analysis works
Project type and purpose
• This is a full-stack Python application for multi-agent competitor analysis.
• It coordinates multiple “agents” (via Crewai), calls a local LLM (Ollama/Llama 3.1 8B), and external search APIs (SerperDevTool) to fetch, analyze, and present insights.
• It exposes both a CLI (test.py) and a Streamlit-based web UI (app.py).
File structure
• Top-level
• requirements.txt – Python dependencies
• README.md – project overview, setup, and architecture notes
• src/
• app.py – Streamlit frontend and request handler
• crew.py – agent orchestration layer using Crewai
• test.py – CLI entry point for quick queries
• config/
– agents.yaml – agent role definitions and parameters
– tasks.yaml – workflows and task sequences
• .env (not shown) – environment variables (SERPER_API_KEY, OpenAI credentials)
README for additional insights
• Multi-agent orchestration via Crewai
• LLM integration (Ollama or OpenAI)
• External web search tool (SerperDevTool) for data aggregation
• CLI and Streamlit interfaces