rohitbind/meeting-mood-analyzer

# 📝 Meeting Mood Analyzer & Recommendation Agent

An interactive AI tool that analyzes meeting notes using an LLM hosted on Groq to determine:

- The **overall mood** of the meeting
- The **meeting type** (e.g., brainstorming, planning)
- The most appropriate **next action**
- A summarized natural language response

Built with Python, [LangChain](https://github.com/langchain-ai/langchain), [Groq](https://groq.com/), and [Streamlit](https://streamlit.io/), this lightweight app can be deployed locally or on [Hugging Face Spaces](https://huggingface.co/spaces).

---

## 🚀 Features

✅ Uses Groq-hosted LLM (`Gemma2-9b-It`) with `temperature=0` for consistent outputs  
✅ Few-shot prompts for accurate classification  
✅ Full fallbacks using rule-based keyword heuristics  
✅ Streamlit-based UI – fast, responsive, and easy to use  
✅ Includes NLP chains via LangChain: mood detection, type classification, and action recommendation  

---

## 🌐 Live Demo (optional)

Coming soon on Hugging Face Spaces!  
URL: `https://huggingface.co/spaces/<your-username>/meeting-mood-analyzer`

---

## 📁 File Structure

---

## 🔧 Setup Instructions

### ✅ Local Development

1. **Clone the repository**

2. **Create a virtual environment (optional but recommended)**

3. **Install dependencies**

4. **Set up your API key**
Create a `.env` file:

> If using VSCode, you can use the `.env.example` file to structure your keys.

5. **Run the application**

---

### ☁️ Deploy on Hugging Face Spaces

1. **Create a new Space**:  
Go to [huggingface.co/spaces](https://huggingface.co/spaces) → "Create Space" → select **Streamlit** SDK.

2. **Upload the following files:**
   - `app.py`
   - `requirements.txt`
   - `README.md`

3. **Set secret key**
   In the Space settings panel under **“Secrets”**, add:

4. That’s it! Once the build completes, your app will be live.

---

## 🔐 Environment Variables

| Name         | Description                             | Required |
|--------------|-----------------------------------------|----------|
| GROQ_API_KEY | API key from [groq.com](https://groq.com/console) | ✅ Yes    |

---

## 🧠 Prompts Used

### 🟩 Mood Detection Prompt
Classifies the sentiment as: `"Positive"`, `"Neutral"`, `"Frustrated"`, `"Anxious"` or `"Conflict"` based on few-shot examples.

### 🟨 Meeting Type Classification
Possible types:
- `Brainstorming`
- `Status Update`
- `Conflict Resolution`
- `Planning`

### 🟥 Next Action Prompt
Suggests the most relevant next step from:
- `Schedule follow-up`
- `Send summary email`
- `Escalate issue`
- `Assign action items`

If LLM fails, deterministic rule-based fallbacks ensure a result is always returned.

---

## 🏗️ Built With

- [LangChain](https://github.com/langchain-ai/langchain)
- [Groq LLMs](https://console.groq.com/)
- [Streamlit](https://streamlit.io/)
- [Python](https://www.python.org/)
- [dotenv](https://pypi.org/project/python-dotenv/)

---

## 📜 License

This project is licensed under the MIT License – feel free to use, adapt, and share!

---

## 🤝 Contributing

Issues and PRs are welcome!  
1. Fork this repo  
2. Create your feature branch (`git checkout -b feature/foo`)  
3. Commit and push  
4. Submit a pull request 🚀

---

## 🧠 Author

Made with ❤️ by [Your Name]  
[LinkedIn](https://linkedin.com/in/yourprofile) • [Portfolio](https://yourportfolio.com)

---

## 🙌 Acknowledgements

Big thanks to:
- [Groq](https://groq.com/)
- [LangChain Community](https://langchain.com/)
- [Streamlit](https://streamlit.io/)
- [Hugging Face](https://huggingface.co)

---

**🎯 Don’t just analyze meetings — understand them.**