🔥 I built this project in just 1 hour to demonstrate how fast agentic AI systems can be prototyped when you focus on modular goal execution, not just LLM completions.

A LangGraph-powered multi-agent system that transforms natural language goals into structured task flows — then executes them using mock tools (like Email, LinkedIn, and Calendar APIs).

“Schedule a product webinar and send invites”

1. 🎯 Goal Parsing – Extracts the intent from user input
2. 💬 Clarification Agent – Asks for more info if the goal is vague
3. 🧠 Fuzzy Task Planner – Uses fuzzy match + keyword logic to map goal to tools
4. 🔧 Tool Executor – Simulates execution (email/calendar/post)
5. 🔁 Feedback Loop – User can say "go again" or "stop"

• LangGraph – Agent workflow orchestration
• Python – CLI backend
• Difflib + input() – Fuzzy matching + interactive user feedback
• Mock Mode – No real APIs needed, perfect for quick demos

• email_sender
• calendar_api
• linkedin_poster
• default_tool (fallback for unknowns)

• Fuzzy match for terms like “webinar”, “invite”, “schedule”
• Falls back to asking the user if the goal is unclear
• Extensible design — easily plug GPT or API tools into each node

• ⏱️ Built in 1 hour — shows execution speed and architectural clarity
• 📦 Agent-based, not just chat-based — structured, multi-step automation
• 🔧 Built for real-world usage — not just a wrapper around OpenAI