How arqamansari23/chatbot-crewai works
Project type and purpose
• This is a lightweight full-stack web application that exposes a chatbot interface via HTML templates, integrates with an LLM (e.g., OpenAI API), persists conversation vectors in a FAISS store, logs leads in a SQLite database, and can send emails via the Mail_Agent module.
• Key indicators:
• Presence of “templates/” and “static/” directories → a server-rendered web UI (likely Flask or similar).
• “Chatbot.py” and “Mail_Agent.py” at root → separate modules for chat logic and mail logic.
• “data/vectorStores/store/index.*” → FAISS vector store for semantic search.
• “leads.db” → SQLite leads storage.
File structure
• Top level:
• Chatbot.py: main chat-server entrypoint
• Mail_Agent.py: email-sending service
• pyproject.toml / requirements.txt: Python dependencies & build config
• leads.db: embedded SQLite for lead data
• Directories:
• data/vectorStores/store: FAISS index files (.faiss, .pkl) for vector embeddings
• static/: CSS/JS assets for the web UI
• templates/: Jinja HTML templates (interface.html, index_MAIL.html)
README
• In this case the README is empty; we infer architecture from code and file layout.
Elements
• Clear labels: “Browser UI,” “Chatbot Server,” “FAISS Store,” “SQLite Leads DB,” “Mail Agent,” “LLM API,” “SMTP”
– Browser → Server (HTTP requests)
– Server → LLM API (HTTP/HTTPS)
– Server ↔ FAISS (local filesystem I/O)
– Server ↔ SQLite (SQL I/O)
– Server → Mail Agent → SMTP (network)