🍽️ Culinary Compass: AI-Powered Recipe Recommender

An intelligent recipe recommender that runs 100% offline using local AI --- no OpenAI API key, no internet, no cost.

Powered by:

• 🧠 crewAI -- for multi-agent orchestration

• 🦙 Ollama + Llama 3 -- for local AI-powered recipe generation

• 🌐 Spoonacular API -- for real recipes (when online)

• 🍲 AI fallback when no recipe is found

Perfect for privacy-conscious users, developers, and food lovers who want smart cooking suggestions --- without sending data to the cloud.

🛠️ How to Run Locally

1. Clone the repo

git clone https://github.com/guruprashanth2004/culinary-compass.git

cd culinary-compass

2. Install dependencies

pip install -r requirements.txt

3. Set up environment variables

Create a .env file in the root folder:

SPOONACULAR_API_KEY=your_spoonacular_key_here

🔑 Get a free key at https://spoonacular.com/food-api

❌ No OpenAI key needed --- powered by Ollama!

4. Run Ollama (Local AI Server)

In a terminal:

ollama run llama3

This downloads the Llama 3 model (~4.7GB) the first time.

5. Launch the Web App

In another terminal:

streamlit run app.py

Open your browser at: http://localhost:8501

---

🔄 How It Works

1. You enter ingredients (e.g., chicken, tomato, onion)

2. Spoonacular API searches for real recipes

3. If no match → AI Recipe Generator (Llama 3) creates a custom recipe

4. Nutritionist Agent adds health insights

5. All powered by crewAI agents working together

✅ Fully offline after setup  

✅ No data leaves your computer  

✅ Fast, private, and intelligent

---

📦 Built With

• crewAI -- Autonomous AI agents

• Ollama -- Run LLMs locally

• Llama 3 -- Meta's open-source large language model

• Streamlit -- Beautiful web interface

• Spoonacular API -- Real recipe data

---

🌟 Features

• ✅ Works offline (AI fallback)

• ✅ No API key needed for AI

• ✅ Private & secure

• ✅ Smart health comments

• ✅ Clean, modern UI

• ✅ Easy to extend

---

📂 Project Structure

recipe_recommender_crewai/

│

├── app.py               # Streamlit web interface

├── crew.py              # RecipeCrew orchestrator

├── agents/

│   ├── ai_recipe_generator.py

│   ├── ingredient_finder.py

│   ├── recipe_suggester.py

│   └── nutritionist.py

├── .env                 # Environment variables

├── .gitignore

├── requirements.txt     # Python dependencies

├── test_crew.py         # Test script

└── README.md            # This file

---

🏆 Why This Stands Out

• 🧠 Uses multi-agent AI to simulate a real culinary team

• 💡 Fallback to local LLM ensures it always works

• 🛡️ No data sent to cloud --- ideal for privacy

• 🎯 Demonstrates real-world AI agent orchestration

---

🌐 Community & Stats

crewAI has powered over 500 million+ agent interactions worldwide.  

🤝 Contributing

Pull requests welcome! Whether it's:

• Adding new agents

• Improving prompts

• UI enhancements

• Diet filters (vegan, keto, etc.)

Let's build the future of AI cooking together.

---

📄 License

MIT License

---

🙌 Made With Love

By Guru Prashanth  

🌐 LinkedIn | 💼 GitHub