README.md

---
title: VoicePulse – Voice Feedback Summarizer 
emoji: 🎧
colorFrom: yellow
colorTo: pink
sdk: gradio
sdk_version: 5.32.1
app_file: app.py
pinned: false
license: mit
short_description: Multilingual audio feedback summarizer
tags:
  - mcp-server-track
  - agent-demo-track
---
# VoicePulse: Multilingual Voice Feedback Collector and Visualizer

VoicePulse is a multilingual voice feedback summarizer built with 🤗 Hugging Face, Gradio, and open-source speech/translation models. Users can speak in Telugu, Hindi, Tamil, Kannada, or English. The app transcribes, translates (if needed), summarizes the feedback, and visualizes key insights using a dynamic word cloud. It also provides CSV export of all collected feedback.

## 🔥 Features

- 🎤 Voice input in 5 languages (English, Telugu, Hindi, Tamil, Kannada)
- 🔁 Auto-translation to English using Facebook NLLB
- ✨ Summarization using DistilBART
- ☁️ Word cloud generation of key terms
- 📁 CSV export of all feedback and summaries
- 🎧 Sample audio playback for testing in all supported languages

## 🚀 How to Use

1. Speak your feedback using your microphone.
2. Choose your spoken language from the dropdown.
3. Click **Process Feedback**.
4. View your summarized feedback and word cloud.
5. Export all collected feedback to CSV if needed.
6. Use sample buttons to try audio feedback in different languages.

## 🛠 Tech Stack

- **Gradio** – UI Interface
- **Hugging Face Transformers** – ASR, translation, and summarization
- **gTTS** – For generating test audio
- **pydub + matplotlib + wordcloud** – Audio trimming and visualization

## 📦 Dependencies
Ensure you have these in your `requirements.txt`:
```txt
gradio
transformers
nltk
wordcloud
matplotlib
pydub
torch
gtts
soundfile
```

## 📄 Tags
`#gradio` `#huggingface` `#voice-feedback` `#multilingual` `#NLLB` `#speech-to-text` `#translation` `#summarization` `#data-viz`

---